提交 5c7a8aee 编写于 作者: H heqiaozhi

merge upstream to my develop

test=develop
Merge remote-tracking branch 'upstream/develop' into develop
...@@ -208,10 +208,10 @@ include(external/xxhash) # download xxhash ...@@ -208,10 +208,10 @@ include(external/xxhash) # download xxhash
include(external/dlpack) include(external/dlpack)
include(external/snappy) # download snappy include(external/snappy) # download snappy
include(external/snappystream) # download snappystream include(external/snappystream) # download snappystream
include(external/warpctc) # download, build, install warpctc
if (NOT WIN32) if (NOT WIN32)
# there is no official support of warpctc, nccl, cupti in windows # there is no official support of nccl, cupti in windows
include(external/warpctc) # download, build, install warpctc
include(cupti) include(cupti)
include(external/gzstream) include(external/gzstream)
endif (NOT WIN32) endif (NOT WIN32)
......
...@@ -18,8 +18,8 @@ ENDIF() ...@@ -18,8 +18,8 @@ ENDIF()
INCLUDE(python_module) INCLUDE(python_module)
FIND_PACKAGE(PythonInterp ${PY_VERSION}) FIND_PACKAGE(PythonInterp ${PY_VERSION} REQUIRED)
FIND_PACKAGE(PythonLibs ${PY_VERSION}) FIND_PACKAGE(PythonLibs ${PY_VERSION} REQUIRED)
if(WIN32) if(WIN32)
execute_process(COMMAND "${PYTHON_EXECUTABLE}" "-c" execute_process(COMMAND "${PYTHON_EXECUTABLE}" "-c"
...@@ -79,6 +79,5 @@ IF(PYTHONINTERP_FOUND) ...@@ -79,6 +79,5 @@ IF(PYTHONINTERP_FOUND)
"please use pip to upgrade protobuf. pip install -U protobuf") "please use pip to upgrade protobuf. pip install -U protobuf")
ENDIF() ENDIF()
ENDIF(PYTHONINTERP_FOUND) ENDIF(PYTHONINTERP_FOUND)
INCLUDE_DIRECTORIES(${PYTHON_INCLUDE_DIR}) INCLUDE_DIRECTORIES(${PYTHON_INCLUDE_DIR})
INCLUDE_DIRECTORIES(${PYTHON_NUMPY_INCLUDE_DIR}) INCLUDE_DIRECTORIES(${PYTHON_NUMPY_INCLUDE_DIR})
...@@ -26,25 +26,33 @@ SET(WARPCTC_INCLUDE_DIR "${WARPCTC_INSTALL_DIR}/include" ...@@ -26,25 +26,33 @@ SET(WARPCTC_INCLUDE_DIR "${WARPCTC_INSTALL_DIR}/include"
# Used in unit test test_WarpCTCLayer # Used in unit test test_WarpCTCLayer
SET(WARPCTC_LIB_DIR "${WARPCTC_INSTALL_DIR}/lib" SET(WARPCTC_LIB_DIR "${WARPCTC_INSTALL_DIR}/lib"
CACHE PATH "Warp-ctc Library Directory" FORCE) CACHE PATH "Warp-ctc Library Directory" FORCE)
SET(WARPCTC_LIBRARIES "${WARPCTC_INSTALL_DIR}/lib/libwarpctc${CMAKE_SHARED_LIBRARY_SUFFIX}"
CACHE FILEPATH "Warp-ctc Library" FORCE)
IF(CMAKE_CXX_COMPILER_ID STREQUAL "Clang" OR CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang" ) IF(CMAKE_CXX_COMPILER_ID STREQUAL "Clang" OR CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang" OR WIN32)
SET(USE_OMP OFF) SET(USE_OMP OFF)
ELSE() ELSE()
SET(USE_OMP ON) SET(USE_OMP ON)
ENDIF() ENDIF()
IF(WIN32)
SET(WARPCTC_REPOSITORY "https://github.com/wopeizl/warp-ctc.git")
ELSE()
SET(WARPCTC_REPOSITORY "https://github.com/dzhwinter/warp-ctc.git")
ENDIF()
ExternalProject_Add( ExternalProject_Add(
extern_warpctc extern_warpctc
${EXTERNAL_PROJECT_LOG_ARGS} ${EXTERNAL_PROJECT_LOG_ARGS}
GIT_REPOSITORY "https://github.com/dzhwinter/warp-ctc.git" GIT_REPOSITORY ${WARPCTC_REPOSITORY}
PREFIX ${WARPCTC_SOURCES_DIR} PREFIX ${WARPCTC_SOURCES_DIR}
UPDATE_COMMAND "" UPDATE_COMMAND ""
CMAKE_ARGS -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER} CMAKE_ARGS -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
-DCMAKE_C_COMPILER=${CMAKE_C_COMPILER} -DCMAKE_C_COMPILER=${CMAKE_C_COMPILER}
-DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS}
-DCMAKE_C_FLAGS=${CMAKE_C_FLAGS} -DCMAKE_C_FLAGS=${CMAKE_C_FLAGS}
-DCMAKE_C_FLAGS_DEBUG=${CMAKE_C_FLAGS_DEBUG}
-DCMAKE_C_FLAGS_RELEASE=${CMAKE_C_FLAGS_RELEASE}
-DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS}
-DCMAKE_CXX_FLAGS_RELEASE=${CMAKE_CXX_FLAGS_RELEASE}
-DCMAKE_CXX_FLAGS_DEBUG=${CMAKE_CXX_FLAGS_DEBUG}
-DCMAKE_INSTALL_PREFIX=${WARPCTC_INSTALL_DIR} -DCMAKE_INSTALL_PREFIX=${WARPCTC_INSTALL_DIR}
-DWITH_GPU=${WITH_GPU} -DWITH_GPU=${WITH_GPU}
-DWITH_OMP=${USE_OMP} -DWITH_OMP=${USE_OMP}
...@@ -59,6 +67,18 @@ ExternalProject_Add( ...@@ -59,6 +67,18 @@ ExternalProject_Add(
-DCMAKE_POSITION_INDEPENDENT_CODE:BOOL=ON -DCMAKE_POSITION_INDEPENDENT_CODE:BOOL=ON
-DCMAKE_INSTALL_PREFIX:PATH=${WARPCTC_INSTALL_DIR} -DCMAKE_INSTALL_PREFIX:PATH=${WARPCTC_INSTALL_DIR}
) )
IF(WIN32)
IF(NOT EXISTS "${WARPCTC_INSTALL_DIR}/lib/warpctc${CMAKE_SHARED_LIBRARY_SUFFIX}")
add_custom_command(TARGET extern_warpctc POST_BUILD
COMMAND cmake -E copy ${WARPCTC_INSTALL_DIR}/bin/warpctc${CMAKE_SHARED_LIBRARY_SUFFIX} ${WARPCTC_INSTALL_DIR}/lib/warpctc${CMAKE_SHARED_LIBRARY_SUFFIX}
)
ENDIF()
SET(WARPCTC_LIBRARIES "${WARPCTC_INSTALL_DIR}/lib/warpctc${CMAKE_SHARED_LIBRARY_SUFFIX}"
CACHE FILEPATH "Warp-ctc Library" FORCE)
else(WIN32)
SET(WARPCTC_LIBRARIES "${WARPCTC_INSTALL_DIR}/lib/libwarpctc${CMAKE_SHARED_LIBRARY_SUFFIX}"
CACHE FILEPATH "Warp-ctc Library" FORCE)
ENDIF(WIN32)
MESSAGE(STATUS "warp-ctc library: ${WARPCTC_LIBRARIES}") MESSAGE(STATUS "warp-ctc library: ${WARPCTC_LIBRARIES}")
INCLUDE_DIRECTORIES(${WARPCTC_INCLUDE_DIR}) # For warpctc code to include its headers. INCLUDE_DIRECTORIES(${WARPCTC_INCLUDE_DIR}) # For warpctc code to include its headers.
......
...@@ -200,6 +200,13 @@ if (WITH_ANAKIN AND WITH_MKL) ...@@ -200,6 +200,13 @@ if (WITH_ANAKIN AND WITH_MKL)
list(APPEND inference_deps anakin_inference_lib) list(APPEND inference_deps anakin_inference_lib)
endif () endif ()
if (TENSORRT_FOUND)
copy(tensorrt_lib DEPS ${inference_deps}
SRCS ${TENSORRT_ROOT}/include/Nv*.h ${TENSORRT_ROOT}/lib/libnvinfer*
DSTS ${FLUID_INSTALL_DIR}/third_party/install/tensorrt/include ${FLUID_INSTALL_DIR}/third_party/install/tensorrt/lib)
endif ()
set(module "inference") set(module "inference")
if(WIN32) if(WIN32)
set(paddle_fluid_lib ${PADDLE_BINARY_DIR}/paddle/fluid/inference/${CMAKE_BUILD_TYPE}/libpaddle_fluid.*) set(paddle_fluid_lib ${PADDLE_BINARY_DIR}/paddle/fluid/inference/${CMAKE_BUILD_TYPE}/libpaddle_fluid.*)
......
...@@ -84,7 +84,7 @@ function(op_library TARGET) ...@@ -84,7 +84,7 @@ function(op_library TARGET)
endif() endif()
if (WIN32) if (WIN32)
# remove windows unsupported op, because windows has no nccl, no warpctc such ops. # remove windows unsupported op, because windows has no nccl, no warpctc such ops.
foreach(windows_unsupport_op "nccl_op" "gen_nccl_id_op" "warpctc_op") foreach(windows_unsupport_op "nccl_op" "gen_nccl_id_op")
if ("${TARGET}" STREQUAL "${windows_unsupport_op}") if ("${TARGET}" STREQUAL "${windows_unsupport_op}")
return() return()
endif() endif()
......
...@@ -209,6 +209,7 @@ paddle.fluid.layers.merge_selected_rows ArgSpec(args=['x', 'name'], varargs=None ...@@ -209,6 +209,7 @@ paddle.fluid.layers.merge_selected_rows ArgSpec(args=['x', 'name'], varargs=None
paddle.fluid.layers.get_tensor_from_selected_rows ArgSpec(args=['x', 'name'], varargs=None, keywords=None, defaults=(None,)) paddle.fluid.layers.get_tensor_from_selected_rows ArgSpec(args=['x', 'name'], varargs=None, keywords=None, defaults=(None,))
paddle.fluid.layers.lstm ArgSpec(args=['input', 'init_h', 'init_c', 'max_len', 'hidden_size', 'num_layers', 'dropout_prob', 'is_bidirec', 'is_test', 'name', 'default_initializer', 'seed'], varargs=None, keywords=None, defaults=(0.0, False, False, None, None, -1)) paddle.fluid.layers.lstm ArgSpec(args=['input', 'init_h', 'init_c', 'max_len', 'hidden_size', 'num_layers', 'dropout_prob', 'is_bidirec', 'is_test', 'name', 'default_initializer', 'seed'], varargs=None, keywords=None, defaults=(0.0, False, False, None, None, -1))
paddle.fluid.layers.psroi_pool ArgSpec(args=['input', 'rois', 'output_channels', 'spatial_scale', 'pooled_height', 'pooled_width', 'name'], varargs=None, keywords=None, defaults=(None,)) paddle.fluid.layers.psroi_pool ArgSpec(args=['input', 'rois', 'output_channels', 'spatial_scale', 'pooled_height', 'pooled_width', 'name'], varargs=None, keywords=None, defaults=(None,))
paddle.fluid.layers.huber_loss ArgSpec(args=['input', 'label', 'delta'], varargs=None, keywords=None, defaults=None)
paddle.fluid.layers.data ArgSpec(args=['name', 'shape', 'append_batch_size', 'dtype', 'lod_level', 'type', 'stop_gradient'], varargs=None, keywords=None, defaults=(True, 'float32', 0, VarType.LOD_TENSOR, True)) paddle.fluid.layers.data ArgSpec(args=['name', 'shape', 'append_batch_size', 'dtype', 'lod_level', 'type', 'stop_gradient'], varargs=None, keywords=None, defaults=(True, 'float32', 0, VarType.LOD_TENSOR, True))
paddle.fluid.layers.open_files ArgSpec(args=['filenames', 'shapes', 'lod_levels', 'dtypes', 'thread_num', 'buffer_size', 'pass_num', 'is_test'], varargs=None, keywords=None, defaults=(None, None, 1, None)) paddle.fluid.layers.open_files ArgSpec(args=['filenames', 'shapes', 'lod_levels', 'dtypes', 'thread_num', 'buffer_size', 'pass_num', 'is_test'], varargs=None, keywords=None, defaults=(None, None, 1, None))
paddle.fluid.layers.read_file ArgSpec(args=['reader'], varargs=None, keywords=None, defaults=None) paddle.fluid.layers.read_file ArgSpec(args=['reader'], varargs=None, keywords=None, defaults=None)
...@@ -375,7 +376,7 @@ paddle.fluid.optimizer.MomentumOptimizer.__init__ ArgSpec(args=['self', 'learnin ...@@ -375,7 +376,7 @@ paddle.fluid.optimizer.MomentumOptimizer.__init__ ArgSpec(args=['self', 'learnin
paddle.fluid.optimizer.MomentumOptimizer.minimize ArgSpec(args=['self', 'loss', 'startup_program', 'parameter_list', 'no_grad_set'], varargs=None, keywords=None, defaults=(None, None, None)) paddle.fluid.optimizer.MomentumOptimizer.minimize ArgSpec(args=['self', 'loss', 'startup_program', 'parameter_list', 'no_grad_set'], varargs=None, keywords=None, defaults=(None, None, None))
paddle.fluid.optimizer.AdagradOptimizer.__init__ ArgSpec(args=['self', 'learning_rate', 'epsilon', 'regularization', 'name'], varargs=None, keywords=None, defaults=(1e-06, None, None)) paddle.fluid.optimizer.AdagradOptimizer.__init__ ArgSpec(args=['self', 'learning_rate', 'epsilon', 'regularization', 'name'], varargs=None, keywords=None, defaults=(1e-06, None, None))
paddle.fluid.optimizer.AdagradOptimizer.minimize ArgSpec(args=['self', 'loss', 'startup_program', 'parameter_list', 'no_grad_set'], varargs=None, keywords=None, defaults=(None, None, None)) paddle.fluid.optimizer.AdagradOptimizer.minimize ArgSpec(args=['self', 'loss', 'startup_program', 'parameter_list', 'no_grad_set'], varargs=None, keywords=None, defaults=(None, None, None))
paddle.fluid.optimizer.AdamOptimizer.__init__ ArgSpec(args=['self', 'learning_rate', 'beta1', 'beta2', 'epsilon', 'regularization', 'name'], varargs=None, keywords=None, defaults=(0.001, 0.9, 0.999, 1e-08, None, None)) paddle.fluid.optimizer.AdamOptimizer.__init__ ArgSpec(args=['self', 'learning_rate', 'beta1', 'beta2', 'epsilon', 'regularization', 'name', 'lazy_mode'], varargs=None, keywords=None, defaults=(0.001, 0.9, 0.999, 1e-08, None, None, False))
paddle.fluid.optimizer.AdamOptimizer.minimize ArgSpec(args=['self', 'loss', 'startup_program', 'parameter_list', 'no_grad_set'], varargs=None, keywords=None, defaults=(None, None, None)) paddle.fluid.optimizer.AdamOptimizer.minimize ArgSpec(args=['self', 'loss', 'startup_program', 'parameter_list', 'no_grad_set'], varargs=None, keywords=None, defaults=(None, None, None))
paddle.fluid.optimizer.AdamaxOptimizer.__init__ ArgSpec(args=['self', 'learning_rate', 'beta1', 'beta2', 'epsilon', 'regularization', 'name'], varargs=None, keywords=None, defaults=(0.001, 0.9, 0.999, 1e-08, None, None)) paddle.fluid.optimizer.AdamaxOptimizer.__init__ ArgSpec(args=['self', 'learning_rate', 'beta1', 'beta2', 'epsilon', 'regularization', 'name'], varargs=None, keywords=None, defaults=(0.001, 0.9, 0.999, 1e-08, None, None))
paddle.fluid.optimizer.AdamaxOptimizer.minimize ArgSpec(args=['self', 'loss', 'startup_program', 'parameter_list', 'no_grad_set'], varargs=None, keywords=None, defaults=(None, None, None)) paddle.fluid.optimizer.AdamaxOptimizer.minimize ArgSpec(args=['self', 'loss', 'startup_program', 'parameter_list', 'no_grad_set'], varargs=None, keywords=None, defaults=(None, None, None))
......
...@@ -50,8 +50,10 @@ cc_library(data_balance_op_handle SRCS data_balance_op_handle.cc DEPS op_handle_ ...@@ -50,8 +50,10 @@ cc_library(data_balance_op_handle SRCS data_balance_op_handle.cc DEPS op_handle_
cc_library(gather_op_handle SRCS gather_op_handle.cc DEPS op_handle_base scope ddim memory variable_visitor) cc_library(gather_op_handle SRCS gather_op_handle.cc DEPS op_handle_base scope ddim memory variable_visitor)
cc_library(fuse_vars_op_handle SRCS fuse_vars_op_handle.cc DEPS op_handle_base scope) cc_library(fuse_vars_op_handle SRCS fuse_vars_op_handle.cc DEPS op_handle_base scope)
cc_library(memory_optimize_pass SRCS analysis_var_pass.cc memory_reuse_types.cc DEPS graph graph_helper pass)
cc_library(modify_op_lock_and_record_event_pass SRCS modify_op_lock_and_record_event_pass.cc DEPS computation_op_handle op_graph_view multi_devices_helper) cc_library(modify_op_lock_and_record_event_pass SRCS modify_op_lock_and_record_event_pass.cc DEPS computation_op_handle op_graph_view multi_devices_helper)
cc_library(memory_early_delete_pass SRCS memory_early_delete_pass.cc DEPS memory_optimize_pass computation_op_handle scale_loss_grad_op_handle rpc_op_handle
all_reduce_op_handle reduce_op_handle broadcast_op_handle data_balance_op_handle graph graph_helper pass)
cc_library(reference_count_pass_helper SRCS reference_count_pass_helper.cc DEPS garbage_collector computation_op_handle) cc_library(reference_count_pass_helper SRCS reference_count_pass_helper.cc DEPS garbage_collector computation_op_handle)
cc_library(eager_deletion_op_handle SRCS eager_deletion_op_handle.cc DEPS lod_tensor selected_rows reference_count_pass_helper) cc_library(eager_deletion_op_handle SRCS eager_deletion_op_handle.cc DEPS lod_tensor selected_rows reference_count_pass_helper)
cc_library(eager_deletion_pass SRCS eager_deletion_pass.cc DEPS computation_op_handle eager_deletion_op_handle graph graph_helper pass) cc_library(eager_deletion_pass SRCS eager_deletion_pass.cc DEPS computation_op_handle eager_deletion_op_handle graph graph_helper pass)
...@@ -63,7 +65,12 @@ cc_library(all_reduce_deps_pass SRCS all_reduce_deps_pass.cc DEPS graph graph_he ...@@ -63,7 +65,12 @@ cc_library(all_reduce_deps_pass SRCS all_reduce_deps_pass.cc DEPS graph graph_he
cc_library(multi_devices_graph_pass SRCS multi_devices_graph_pass.cc DEPS multi_devices_helper computation_op_handle cc_library(multi_devices_graph_pass SRCS multi_devices_graph_pass.cc DEPS multi_devices_helper computation_op_handle
scale_loss_grad_op_handle rpc_op_handle all_reduce_op_handle reduce_op_handle broadcast_op_handle data_balance_op_handle fused_broadcast_op_handle) scale_loss_grad_op_handle rpc_op_handle all_reduce_op_handle reduce_op_handle broadcast_op_handle data_balance_op_handle fused_broadcast_op_handle)
set(SSA_GRAPH_EXECUTOR_DEPS graph framework_proto sequential_execution_pass modify_op_lock_and_record_event_pass all_reduce_deps_pass reference_count_pass eager_deletion_pass) set(SSA_GRAPH_EXECUTOR_DEPS graph framework_proto sequential_execution_pass modify_op_lock_and_record_event_pass all_reduce_deps_pass reference_count_pass eager_deletion_pass memory_optimize_pass memory_early_delete_pass)
if (WITH_GPU)
list(APPEND SSA_GRAPH_EXECUTOR_DEPS reference_count_pass)
endif()
cc_test(memory_reuse_types_test SRCS memory_reuse_types_test.cc memory_reuse_types.cc DEPS framework_proto graph)
cc_test(analysis_var_pass_test SRCS analysis_var_pass_test.cc analysis_var_pass.cc memory_reuse_types.cc DEPS framework_proto graph graph_helper op_registry pass)
cc_library(ssa_graph_executor SRCS ssa_graph_executor.cc DEPS ${SSA_GRAPH_EXECUTOR_DEPS}) cc_library(ssa_graph_executor SRCS ssa_graph_executor.cc DEPS ${SSA_GRAPH_EXECUTOR_DEPS})
...@@ -84,4 +91,5 @@ cc_test(fused_broadcast_op_test SRCS fused_broadcast_op_handle_test.cc DEPS fuse ...@@ -84,4 +91,5 @@ cc_test(fused_broadcast_op_test SRCS fused_broadcast_op_handle_test.cc DEPS fuse
cc_library(build_strategy SRCS build_strategy.cc DEPS cc_library(build_strategy SRCS build_strategy.cc DEPS
graph_viz_pass multi_devices_graph_pass graph_viz_pass multi_devices_graph_pass
multi_devices_graph_print_pass multi_devices_graph_check_pass multi_devices_graph_print_pass multi_devices_graph_check_pass
fuse_elewise_add_act_pass multi_batch_merge_pass) fuse_elewise_add_act_pass multi_batch_merge_pass
memory_optimize_pass)
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#include "paddle/fluid/framework/details/analysis_var_pass.h"
#include <algorithm>
#include <atomic>
#include <deque>
#include <fstream>
#include <iostream>
#include <iterator>
#include <memory>
#include <queue>
#include <sstream>
#include <string>
#include <type_traits>
#include <vector>
#include "gflags/gflags.h"
#include "paddle/fluid/framework/data_type.h"
#include "paddle/fluid/framework/ir/graph.h"
#include "paddle/fluid/framework/ir/graph_helper.h"
DEFINE_bool(enable_subgraph_optimize, false,
"SubGraph also reuse global graph variables, it will reduce the "
"memory occupation"
"but a higher risk of memory reuse error. default disabled.");
DEFINE_string(memory_optimize_debug, "",
"debug the operator output variable when do the variable reuse."
"memory reuse pass."
"only for debug, default disabled.");
namespace paddle {
namespace framework {
namespace details {
static inline bool IsSameDesc(OpDesc* op1, OpDesc* op2) {
return op1->Type() == op2->Type() && op1->Inputs() == op2->Inputs() &&
op1->Outputs() == op2->Outputs();
}
template <typename Container, typename Callback>
class FilterVariableImpl {
public:
void operator()(const Container& nodes, Callback callback) {
for (auto* node : nodes) {
callback(node);
}
}
};
// filter var node for op->inputs/outputs
template <typename Callback>
class FilterVariableImpl<std::vector<ir::Node*>, Callback> {
public:
void operator()(const std::vector<ir::Node*>& nodes, Callback callback) {
for (auto* var : nodes) {
if (var->IsVar() && !var->IsCtrlVar()) {
callback(var);
}
}
}
};
template <typename Container, typename Callback>
void FilterVariables(const Container& nodes, Callback callback) {
FilterVariableImpl<Container, Callback>()(nodes, callback);
}
std::unique_ptr<ir::Graph> AnalysisVarPass::ApplyImpl(
std::unique_ptr<ir::Graph> graph) const {
auto nodes = graph->Nodes();
auto subblock_vars = GetSubBlockVars(nodes);
skip_set_.insert(subblock_vars.begin(), subblock_vars.end());
cfg_.reset(new details::ControlFlowGraph(*graph));
cfg_->LiveVariableAnalysis();
InitSSAGraphNodes();
int reuse_id = 0;
for (size_t idx = 0; idx < cfg_->Ops().size(); ++idx) {
auto& op = cfg_->Ops()[idx];
auto* op_desc = op->Op();
// some op in graph has no op desc
if (op_desc == nullptr) continue;
if (OpHasSubBlock(op_desc)) {
if (FLAGS_enable_subgraph_optimize) {
SubGraphOptimize(op_desc);
} else {
VLOG(3) << op->Name()
<< " has subblock, but disable subgraph optimize. skipped.";
continue;
}
}
for (auto& var : op->outputs) {
if (NodeCanReused(var) && cfg_->Use(op).count(var->Name()) == 0) {
ir::Node* cache = pool_.NodeMatch(var);
if (var->Name() == FLAGS_memory_optimize_debug) {
VLOG(3) << "start match var " << DebugString(var) << " of op "
<< op->Name();
VLOG(3) << pool_.ToString();
VLOG(3) << "matched in pool : "
<< ((cache == nullptr) ? "False" : "True");
}
if (cache != nullptr) {
if (var->Name() == cache->Name()) {
VLOG(3) << "The same cache variable is cascade reused."
<< var->Name() << " is re-filled to the pool after"
<< "the reused op is finished. Current op can not "
<< "replace it again. Skip this candidate.";
continue;
}
int node_idx_in_pool = pool_.GetIndex(cache);
VLOG(3) << string::Sprintf(
"!!! %s, %s => %s, cache idx %d, pool size %d",
std::to_string(reuse_id++), DebugString(var), DebugString(cache),
node_idx_in_pool, static_cast<int>(pool_.size()));
// update CFG Graph on the fly.
// reused var maybe re-fill into the pool
cfg_->RenameVarInCFGGraph(var->Name(), cache->Name(), idx);
// NOTE(dzhwinter): we need to both update the ProgramDesc
// and IR Graph. because op_desc/var_desc is used in CreateOp,
// CreateVar when running happens. But IR Graph
// define the dependence relationship between nodes.
RenameVarInGraphDesc(var->Name(), cache->Name(), idx);
RenameVarInGraphNode(var->Name(), cache->Name(), idx, graph.get());
pool_.Erase(cache);
}
}
}
// fill the pool
for (auto var : cfg_->LiveIn(op)) {
if (cfg_->LiveOut(op).count(var) == 0) {
ir::Node* var_node = cfg_->GetNodeFromVarName(var, op);
if (var_node == nullptr) continue;
if (NodeCanReused(var_node) && !pool_.Has(var_node)) {
pool_.Insert(var_node, op);
}
}
}
}
graph->ResolveHazard(var_nodes_);
// For early delete pass. use GraphNodePool load the unlived vars.
// 1. find all deps op for each unlived var in memory pool.
for (auto& op : graph->Nodes()) {
for (auto& var : op->inputs) {
if (pool_.Has(var)) {
pool_.Insert(var, op);
}
}
}
// 2. convert ir node based memory pool to graph node
// because Node* maybe released bettwen passes.
auto& graph_pool = graph->Get<GraphNodePool>(kGraphNodePool);
for (auto it = pool_.begin(); it != pool_.end(); ++it) {
std::unordered_set<OpDesc*> descs;
for (auto& op : it->second) {
PADDLE_ENFORCE(op->IsOp());
descs.insert(op->Op());
}
graph_pool.push_back(std::make_pair(it->first->Name(), descs));
}
return graph;
}
void AnalysisVarPass::SubGraphOptimize(OpDesc* op_desc) const {
// conditional block, while op and their grad op
auto* sub_block_desc =
AttrReader(op_desc->GetAttrMap()).Get<BlockDesc*>("sub_block");
// create a mirror block to construct an IR Graph.
ProgramDesc prog;
auto* copy_block = prog.MutableBlock(0);
for (auto* op : sub_block_desc->AllOps()) {
auto* copy_op = copy_block->AppendOp();
copy_op->CopyFrom(*op);
copy_op->Flush();
}
for (auto* var : sub_block_desc->AllVars()) {
auto* copy_var = copy_block->Var(var->Name());
copy_var->SetDataType(var->GetDataType());
// only lod tensor can be reused. So ignore the multiple dims case.
copy_var->SetType(var->GetType());
copy_var->SetShape(var->GetShape());
copy_var->SetPersistable(var->Persistable());
}
ir::Graph sub_graph(prog);
std::unordered_set<ir::Node*> sub_graph_all_ops;
FilterVariables(sub_graph.Nodes(), [&](ir::Node* var) {
// sub_graph_all_ops.emplace(var);
if (var->IsVar() && !var->IsCtrlVar()) {
sub_graph_all_ops.emplace(var);
}
});
int sub_reuse_id = 0;
// subgraph nodes is unordered, reuse need to follow the desc order.
// find the right op node through the descs
for (auto* sub_op_desc : sub_block_desc->AllOps()) {
ir::Node* sub_op = nullptr;
for (auto* node : sub_graph_all_ops) {
if (node->Op() == sub_op_desc) {
sub_op = node;
break;
}
}
PADDLE_ENFORCE(sub_op != nullptr);
for (auto* var : sub_op->outputs) {
if (NodeCanReused(var)) {
ir::Node* cache = pool_.NodeMatch(var);
if (cache != nullptr) {
if (var->Var()->GetDataType() != cache->Var()->GetDataType()) {
continue;
}
int node_idx_in_pool = pool_.GetIndex(cache);
VLOG(3) << string::Sprintf(
"!!! %s, %s => %s, cache idx %d, pool size %d",
std::to_string(sub_reuse_id++), DebugString(var),
DebugString(cache), node_idx_in_pool,
static_cast<int>(pool_.size()));
// NOTE(dzh): subblock is not in IR graph. Modify the block_desc
// immediately to make the subblock variable reuse strategy take
// effect. Because it is a single op in graph. No need to
// update the ir nodes.
sub_op_desc->Rename(var->Name(), cache->Name());
if (sub_op_desc->Block()->HasVar(var->Name())) {
sub_op_desc->Block()->RemoveVar(var->Name());
}
}
}
}
}
}
std::unordered_set<std::string> AnalysisVarPass::GetSubBlockVars(
const std::unordered_set<ir::Node*>& nodes) const {
std::unordered_set<std::string> vars;
for (auto& op : nodes) {
if (!op->IsOp() || op->Op() == nullptr) continue;
auto* op_desc = op->Op();
if (OpHasSubBlock(op_desc)) {
auto inputs = op_desc->InputArgumentNames();
auto outputs = op_desc->OutputArgumentNames();
vars.insert(inputs.begin(), inputs.end());
vars.insert(outputs.begin(), outputs.end());
}
}
return vars;
}
void AnalysisVarPass::RenameVarInGraphDesc(const std::string& var,
const std::string& cache_var,
size_t idx) const {
for (size_t i = idx; i < cfg_->Ops().size(); ++i) {
auto* op = cfg_->Ops()[i];
PADDLE_ENFORCE(op->IsOp() && op->Op());
auto* op_desc = op->Op();
op_desc->RenameInput(var, cache_var);
op_desc->RenameOutput(var, cache_var);
if (op_desc->Block()->HasVar(var)) op_desc->Block()->RemoveVar(var);
op_desc->Flush();
}
}
void AnalysisVarPass::InitSSAGraphNodes() const {
std::unordered_map<std::string, std::unordered_set<ir::Node*>> all_vars;
if (var_nodes_.empty()) {
for (auto* op : cfg_->Ops()) {
for (auto* node : op->inputs) {
if (all_vars[node->Name()].count(node) == 0) {
all_vars[node->Name()].emplace(node);
var_nodes_[node->Name()].emplace_back(node);
}
}
for (auto* node : op->outputs) {
if (all_vars[node->Name()].count(node) == 0) {
all_vars[node->Name()].emplace(node);
var_nodes_[node->Name()].emplace_back(node);
}
}
}
}
}
void AnalysisVarPass::RenameVarInGraphNode(const std::string& var,
const std::string& cache_var,
size_t idx, ir::Graph* graph) const {
// if replace happens, we need to create a newer version cache_var
// but use the same dims/data_type with var.
PADDLE_ENFORCE(var_nodes_[var].size() >= 1 &&
var_nodes_[var].at(0)->Var() != nullptr);
std::unique_ptr<VarDesc> var_desc(new VarDesc(*var_nodes_[var].at(0)->Var()));
var_desc->SetName(cache_var);
for (size_t i = idx; i < cfg_->Ops().size(); ++i) {
auto* op = cfg_->Ops()[i];
// redirect the input to the latest version of cache_var
for (auto* node : op->inputs) {
if (node->Name() == var) {
ir::Node* cache_node = graph->CreateVarNode(var_desc.get());
var_nodes_[cache_var].emplace_back(cache_node);
// swap node to cache_node
cache_node->outputs.insert(cache_node->outputs.end(),
node->outputs.begin(), node->outputs.end());
PADDLE_ENFORCE(node->inputs.size() == 1 && node->inputs[0]->IsOp());
auto* prev_op = node->inputs[0];
std::replace(prev_op->outputs.begin(), prev_op->outputs.end(), node,
cache_node);
cache_node->inputs.emplace_back(prev_op);
for (auto* next_op : node->outputs) {
std::replace(next_op->inputs.begin(), next_op->inputs.end(), node,
cache_node);
}
}
}
// if we need to rename the output,
// always create a newer version of cache_var
for (auto* node : op->outputs) {
if (node->Name() == var) {
ir::Node* cache_node = graph->CreateVarNode(var_desc.get());
var_nodes_[cache_var].emplace_back(cache_node);
// swap node to cache node
cache_node->outputs.insert(cache_node->outputs.end(),
node->outputs.begin(), node->outputs.end());
cache_node->inputs.emplace_back(op);
std::replace(op->outputs.begin(), op->outputs.end(), node, cache_node);
for (auto* next_op : node->outputs) {
std::replace(next_op->inputs.begin(), next_op->inputs.end(), node,
cache_node);
}
}
}
}
// release node of unused var in graph
for (auto* node : var_nodes_[var]) {
graph->RemoveNode(node);
}
var_nodes_.at(var).clear();
}
bool AnalysisVarPass::NodeCanReused(ir::Node* node) const {
if (!node->IsVar() || node->IsCtrlVar()) return false;
auto* desc = node->Var();
auto type = desc->GetType();
if (desc->Persistable() || type != proto::VarType::LOD_TENSOR ||
desc->GetShape().empty()) {
return false;
}
// vars can be @EMPTY@, @LR_DECAY_REUSE_ID@. For example, while_grad
std::string name = node->Name();
if (!name.empty() && name[0] == '@' && name[name.size() - 1] == '@')
return false;
if (skip_set_.count(name)) return false;
for (auto* op : node->inputs) {
if (op->Op()->HasAttr("force_cpu")) {
// op output force generated in cpu, can not be reused.
return framework::AttrReader(op->Op()->GetAttrMap())
.Get<bool>("force_cpu") == 0;
}
}
return true;
}
bool AnalysisVarPass::OpHasSubBlock(OpDesc* desc) const {
const AttributeMap& attrs = desc->GetAttrMap();
for (auto& attr : attrs) {
if (attr.second.type() == typeid(BlockDesc*) || // NOLINT
attr.second.type() == typeid(std::vector<BlockDesc*>)) // NOLINT
return true;
}
return false;
}
std::vector<ir::Node*> SortOpLikeDescOrder(const ir::Graph& graph) {
PADDLE_ENFORCE(graph.Has(kAllOpDescs),
"Graph has no attribute of kAllOpDescs.");
// 1. get op desc order
auto& op_descs = graph.Get<const std::vector<OpDesc*>>(kAllOpDescs);
// 2. topology sort order
auto nodes = graph.Nodes();
std::deque<ir::Node*> ops;
FilterVariables(nodes, [&](ir::Node* op) {
if (op->IsOp() && op->Op() != nullptr) {
ops.emplace_back(op);
}
});
std::unordered_map<ir::Node*, size_t> op_deps;
std::list<ir::Node*> ready_ops;
std::unordered_map<ir::Node*, std::unordered_set<ir::Node*>> pending_ops;
for (auto* op : ops) {
std::unordered_set<ir::Node*> preceding_op;
for (auto* in : op->inputs) {
if (in->inputs.empty()) continue;
PADDLE_ENFORCE(in->inputs.size() == 1 && in->inputs[0]->IsOp());
preceding_op.emplace(in->inputs[0]);
pending_ops[in->inputs[0]].emplace(op);
}
op_deps[op] = preceding_op.size();
if (preceding_op.empty()) {
ready_ops.emplace_back(op);
}
}
// 3. generated op list based desc order and the topology order
std::vector<ir::Node*> ret;
std::list<OpDesc*> op_descs_list(op_descs.begin(), op_descs.end());
auto update_by_found_node = [&](ir::Node* found_node) {
for (auto* pending_op : pending_ops[found_node]) {
if (--op_deps[pending_op] == 0) {
ready_ops.emplace_back(pending_op);
}
}
ready_ops.remove(found_node);
ret.emplace_back(found_node);
};
while (!ready_ops.empty()) {
bool all_of_ready_op_unmatched = true;
for (auto it = op_descs_list.begin(); it != op_descs_list.end();) {
auto op_desc = *it;
ir::Node* found_node = nullptr;
for (auto* op : ready_ops) {
if (IsSameDesc(op->Op(), op_desc)) {
found_node = op;
break;
}
}
// 3.1 op desc deleted by other pass
if (found_node == nullptr) {
++it;
continue;
} else {
all_of_ready_op_unmatched = false;
it = op_descs_list.erase(it);
}
update_by_found_node(found_node);
}
// 3.2 op descs are added by other pass
// preceding op non empty means some new op descs are
// created, but not contained in return node list.
// these new op desc may depend on each other.
std::list<ir::Node*> prev_ready_ops(ready_ops);
if (all_of_ready_op_unmatched) {
for (auto op : prev_ready_ops) {
update_by_found_node(op);
}
}
}
PADDLE_ENFORCE(std::all_of(
op_deps.begin(), op_deps.end(),
[&](const std::pair<ir::Node*, size_t>& p) { return p.second == 0; }));
return ret;
}
ControlFlowGraph::ControlFlowGraph(const ir::Graph& graph) {
ops_ = SortOpLikeDescOrder(graph);
ConnectNodes();
}
void ControlFlowGraph::BuildCFGGraph() {
// FIXME(dzh): same effect with ConnectNodes, but use the control
// link to build dependency graph, it goes wrong in transformer.
for (ir::Node* op : ops_) {
for (auto& input_var : op->inputs) {
if (!input_var->inputs.empty()) {
PADDLE_ENFORCE(
input_var->inputs.size() == 1 && input_var->inputs[0]->IsOp(),
"Preceding Op Node of Var Node must be unique");
auto* pred_op = input_var->inputs[0];
if (pred_op->Op() != nullptr) {
predecessors_[op].insert(pred_op);
successors_[pred_op].insert(op);
}
}
if (input_var->IsVar() && !input_var->IsCtrlVar()) {
uses_[op].insert(input_var->Name());
}
}
for (auto& output_var : op->outputs) {
// output var may be used by many op
for (auto* succ_op : output_var->outputs) {
if (succ_op->Op() != nullptr) {
successors_[op].insert(succ_op);
predecessors_[succ_op].insert(op);
}
}
if (output_var->IsVar() && !output_var->IsCtrlVar()) {
defs_[op].insert(output_var->Name());
}
}
}
}
void ControlFlowGraph::ConnectNodes() {
for (size_t i = 0; i < ops_.size(); ++i) {
auto& op = ops_[i];
try {
auto& next_op = ops_.at(i + 1);
successors_[op].insert(next_op);
predecessors_[next_op].insert(op);
} catch (...) {
// do nothing
}
FilterVariables(op->inputs,
[&](ir::Node* var) { uses_[op].emplace(var->Name()); });
FilterVariables(op->outputs,
[&](ir::Node* var) { defs_[op].emplace(var->Name()); });
}
}
void ControlFlowGraph::LiveVariableAnalysis() {
// NOTE(dzh): variable liveless analysis (a.k.a reversed_ops algorithm)
// compute the liveness of for each variable though reversed_ops algorithm.
// It iterates the operators from end to begin, compute the live in/live out
// variable set for each op, then the diff between in/out will be used for
// the variable reuse. For detail refer to
// http://www.cs.cornell.edu/courses/cs4120/2013fa/lectures/lec26-fa13.pdf
std::list<ir::Node*> work_list(ops_.rbegin(), ops_.rend());
while (!work_list.empty()) {
ir::Node* op = work_list.front();
work_list.pop_front();
// get the live_in calculated before. Empty if first.
auto prev_live_in = std::move(live_in_[op]);
for (auto& s : successors_[op]) {
for (auto& var : live_in_[s]) {
live_out_[op].insert(var);
}
}
for (auto& var : uses_[op]) {
live_in_[op].insert(var);
}
for (auto& var : live_out_[op]) {
live_in_[op].insert(var);
}
for (auto& var : defs_[op]) {
live_in_[op].erase(var);
}
// If the live_in is not changed, then the liveness analysis of
// predecessors is completed.
//
// Otherwise, recalculate the predecessors liveness
if (live_in_[op] != prev_live_in) {
for (auto& pre : predecessors_[op]) {
work_list.push_back(pre);
}
}
}
}
void ControlFlowGraph::RenameVarInCFGGraph(const std::string& old_node,
const std::string& new_node,
int begin_idx) {
// update graph from begin idx to the end
for (size_t i = begin_idx; i != ops_.size(); ++i) {
auto* op = ops_[i];
if (uses_[op].find(old_node) != uses_[op].end()) {
uses_[op].erase(old_node);
uses_[op].insert(new_node);
}
if (defs_[op].find(old_node) != defs_[op].end()) {
defs_[op].erase(old_node);
defs_[op].insert(new_node);
}
if (live_in_[op].find(old_node) != live_in_[op].end()) {
live_in_[op].erase(old_node);
live_in_[op].insert(new_node);
}
if (live_out_[op].find(old_node) != live_out_[op].end()) {
live_out_[op].erase(old_node);
live_out_[op].insert(new_node);
}
}
}
const std::set<std::string> ControlFlowGraph::LiveIn(ir::Node* op) const {
auto it = live_in_.find(op);
PADDLE_ENFORCE(
it != live_in_.end(),
string::Sprintf("Expect %s in live_in, but Not Found.", op->Name()));
return it->second;
}
const std::set<std::string> ControlFlowGraph::LiveOut(ir::Node* op) const {
auto it = live_out_.find(op);
PADDLE_ENFORCE(
it != live_out_.end(),
string::Sprintf("Expect %s in live_out, but Not Found.", op->Name()));
return it->second;
}
const std::set<std::string> ControlFlowGraph::Use(ir::Node* op) const {
auto it = uses_.find(op);
PADDLE_ENFORCE(
it != uses_.end(),
string::Sprintf("Expect %s in live_out, but Not Found.", op->Name()));
return it->second;
}
const std::vector<ir::Node*> ControlFlowGraph::Ops() const { return ops_; }
std::vector<ir::Node*>& ControlFlowGraph::Ops() { return ops_; }
ir::Node* ControlFlowGraph::GetNodeFromVarName(const std::string& name,
ir::Node* op) const {
// in ssa-graph, different version nodes have same name,
// this function get the latest version var before target op
// It may return nullptr, such as data node.
ir::Node* found_node = nullptr;
for (auto* node : ops_) {
if (node == op) break;
for (auto& output : node->outputs) {
if (output->Name() == name) {
found_node = output;
}
}
}
return found_node;
}
} // namespace details
} // namespace framework
} // namespace paddle
REGISTER_PASS(analysis_var_pass, paddle::framework::details::AnalysisVarPass)
.RequireGraphAttr(paddle::framework::details::kGraphNodePool)
.RequireGraphAttr(paddle::framework::details::kAllOpDescs);
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#pragma once
#include <algorithm>
#include <list>
#include <map>
#include <memory>
#include <set>
#include <string>
#include <unordered_map>
#include <utility>
#include <vector>
#include "paddle/fluid/framework/data_type.h"
#include "paddle/fluid/framework/details/memory_reuse_types.h"
#include "paddle/fluid/framework/ir/graph.h"
#include "paddle/fluid/framework/ir/pass.h"
namespace paddle {
namespace framework {
namespace details {
constexpr char kAllOpDescs[] = "all_op_descs";
std::vector<ir::Node*> SortOpLikeDescOrder(const ir::Graph& graph);
// sort op in bfs order
std::vector<ir::Node*> BFSSortGraphOps(const ir::Graph& graph);
class ControlFlowGraph;
class AnalysisVarPass : public ir::Pass {
protected:
std::unique_ptr<ir::Graph> ApplyImpl(
std::unique_ptr<ir::Graph> graph) const override;
private:
// fill the variable map(var_nodes) by version.
void InitSSAGraphNodes() const;
// update program descs
void RenameVarInGraphDesc(const std::string& var,
const std::string& cache_var, size_t idx) const;
// update ir nodes
void RenameVarInGraphNode(const std::string& var,
const std::string& cache_var, size_t idx,
ir::Graph* graph) const;
void SubGraphOptimize(OpDesc* op_desc) const;
// valid a tensor can be reuse or not
bool NodeCanReused(ir::Node* node) const;
// scan subblock and collect the output/input variables.
std::unordered_set<std::string> GetSubBlockVars(
const std::unordered_set<ir::Node*>&) const;
// check op has subblock or not
bool OpHasSubBlock(OpDesc* desc) const;
private:
// Reuse Node Pool, Owned.
mutable OrderedNodePairPool pool_;
// controlflow Graph
mutable std::unique_ptr<ControlFlowGraph> cfg_;
// skip set
mutable std::unordered_set<std::string> skip_set_;
// var nodes
mutable std::map<std::string, std::vector<ir::Node*>> var_nodes_;
};
class ControlFlowGraph {
public:
ControlFlowGraph() = default;
// For IR Graph in parallelexecutor
explicit ControlFlowGraph(const ir::Graph& graph);
void LiveVariableAnalysis();
void RenameVarInCFGGraph(const std::string& old_node,
const std::string& new_node, int begin_idx);
const std::set<std::string> LiveIn(ir::Node* op) const;
const std::set<std::string> LiveOut(ir::Node* op) const;
const std::set<std::string> Use(ir::Node* op) const;
const std::vector<ir::Node*> Ops() const;
std::vector<ir::Node*>& Ops();
// for ssa-graph nodes
ir::Node* GetNodeFromVarName(const std::string& name, ir::Node* op) const;
private:
void BuildCFGGraph();
void ConnectNodes();
using NodeListMap = std::unordered_map<ir::Node*, std::set<ir::Node*>>;
using VarSetMap = std::map<ir::Node*, std::set<std::string>>;
// successors ops use the output variables.
NodeListMap successors_;
// predecessors ops generated input variables.
NodeListMap predecessors_;
// variables lived before run current op.
VarSetMap live_in_;
// variables lived after run current op.
VarSetMap live_out_;
VarSetMap uses_; // op inputs
VarSetMap defs_; // op outputs
std::vector<ir::Node*> ops_; // op sequence by topology sort
};
} // namespace details
} // namespace framework
} // namespace paddle
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#include "paddle/fluid/framework/details/analysis_var_pass.h"
#include <algorithm>
#include <iostream>
#include <iterator>
#include "glog/logging.h"
#include "gtest/gtest.h"
#include "paddle/fluid/framework/ir/graph.h"
#include "paddle/fluid/framework/ir/graph_helper.h"
#include "paddle/fluid/framework/op_registry.h"
#include "paddle/fluid/framework/operator.h"
#include "paddle/fluid/framework/program_desc.h"
namespace paddle {
namespace framework {
class DummyOp : public OperatorBase {
public:
DummyOp(const std::string& type, const VariableNameMap& inputs,
const VariableNameMap& outputs, const AttributeMap& attrs)
: OperatorBase(type, inputs, outputs, attrs) {}
private:
void RunImpl(const Scope& scope,
const platform::Place& place) const override {}
};
class SumOpMaker : public OpProtoAndCheckerMaker {
public:
void Make() {
AddInput("X", "").AsDuplicable();
AddOutput("Out", "");
AddComment("");
}
};
class AssignOpMaker : public OpProtoAndCheckerMaker {
public:
void Make() {
AddInput("X", "").AsDuplicable();
AddOutput("Out", "");
AddComment("");
}
};
class DummyVarTypeInference : public VarTypeInference {
public:
void operator()(const OpDesc& op_desc, BlockDesc* block) const override {
auto& inputs = op_desc.Input("X");
auto type = block->Var(inputs.front())->GetType();
auto out_var_name = op_desc.Output("Out").front();
block->Var(out_var_name)->SetType(type);
}
};
} // namespace framework
} // namespace paddle
REGISTER_OPERATOR(sum, paddle::framework::DummyOp,
paddle::framework::SumOpMaker,
paddle::framework::DummyVarTypeInference);
REGISTER_OPERATOR(assign, paddle::framework::DummyOp,
paddle::framework::AssignOpMaker,
paddle::framework::DummyVarTypeInference);
REGISTER_OPERATOR(dummy, paddle::framework::DummyOp,
paddle::framework::SumOpMaker,
paddle::framework::DummyVarTypeInference);
/*
https://en.wikipedia.org/wiki/Live_variable_analysis
Create a customed classical dependency graph, left row is the instruction
number.
1. a = 1
2. b = a
3. c = a
4. d = b + c
5. e = d
a--------+
| |
b c
| |
d--------+
|
e
Then analysis these variable's liveness range
*/
namespace paddle {
namespace framework {
namespace details {
static inline bool IsSameDesc(OpDesc* op1, OpDesc* op2) {
return op1->Type() == op2->Type() && op1->Inputs() == op2->Inputs() &&
op1->Outputs() == op2->Outputs();
}
inline static ProgramDesc FillProgramDesc() {
ProgramDesc prog;
prog.MutableBlock(0)->Var("a")->SetType(proto::VarType::LOD_TENSOR);
prog.MutableBlock(0)->Var("b")->SetType(proto::VarType::LOD_TENSOR);
prog.MutableBlock(0)->Var("c")->SetType(proto::VarType::LOD_TENSOR);
prog.MutableBlock(0)->Var("d")->SetType(proto::VarType::LOD_TENSOR);
prog.MutableBlock(0)->Var("e")->SetType(proto::VarType::LOD_TENSOR);
{
auto* op = prog.MutableBlock(0)->AppendOp();
op->SetType("assign");
op->SetInput("X", {"a"});
op->SetOutput("Out", {"b"});
}
{
auto* op = prog.MutableBlock(0)->AppendOp();
op->SetType("assign");
op->SetInput("X", {"a"});
op->SetOutput("Out", {"c"});
}
{
auto* op = prog.MutableBlock(0)->AppendOp();
op->SetType("sum");
op->SetInput("X", {"b", "c"});
op->SetOutput("Out", {"d"});
}
{
auto* op = prog.MutableBlock(0)->AppendOp();
op->SetType("assign");
op->SetInput("X", {"d"});
op->SetOutput("Out", {"e"});
}
return prog;
}
template <typename Container>
inline static std::string DebugString(const Container& c) {
std::stringstream ss;
for (auto& item : c) {
ss << item << " ";
}
return ss.str();
}
TEST(CFGGraph, IRGraph) {
// prepare ir graph
auto prog = FillProgramDesc();
ir::Graph graph(prog);
const std::vector<OpDesc*>* all_op_descs =
new std::vector<OpDesc*>(prog.Block(0).AllOps());
graph.Set(details::kAllOpDescs, all_op_descs); // take ownership
ControlFlowGraph cfg(graph);
cfg.LiveVariableAnalysis();
// test assign op
ASSERT_TRUE((std::set<std::string>{"a"} == cfg.LiveIn(cfg.Ops()[0])));
ASSERT_TRUE((std::set<std::string>{"a", "b"} == cfg.LiveOut(cfg.Ops()[0])));
// test assign op
ASSERT_TRUE((std::set<std::string>{"a", "b"} == cfg.LiveIn(cfg.Ops()[1])));
ASSERT_TRUE((std::set<std::string>{"b", "c"} == cfg.LiveOut(cfg.Ops()[1])));
// test sum op
ASSERT_TRUE((std::set<std::string>{"b", "c"} == cfg.LiveIn(cfg.Ops()[2])));
ASSERT_TRUE((std::set<std::string>{"d"} == cfg.LiveOut(cfg.Ops()[2])));
// test assign op
ASSERT_TRUE((std::set<std::string>{"d"} == cfg.LiveIn(cfg.Ops()[3])));
ASSERT_TRUE((std::set<std::string>{} == cfg.LiveOut(cfg.Ops()[3])));
}
// 1. normal test
TEST(SortOpLikeDescOrder, NormalTest) {
auto prog = FillProgramDesc();
ir::Graph graph(prog);
const std::vector<OpDesc*>* all_op_descs =
new std::vector<OpDesc*>(prog.Block(0).AllOps());
graph.Set(details::kAllOpDescs, all_op_descs); // take ownership
auto nodes = SortOpLikeDescOrder(graph);
auto op_descs = prog.Block(0).AllOps();
for (size_t i = 0; i < nodes.size(); ++i) {
auto node = nodes[i];
auto op_desc = op_descs[i];
ASSERT_TRUE(IsSameDesc(node->Op(), op_desc));
}
}
// 2. remove some op_desc
TEST(SortOpLikeDescOrder, RemoveOpDesc) {
auto prog = FillProgramDesc();
ir::Graph graph(prog);
const std::vector<OpDesc*>* all_op_descs =
new std::vector<OpDesc*>(prog.Block(0).AllOps());
graph.Set(details::kAllOpDescs, all_op_descs); // take ownership
auto nodes = graph.Nodes();
auto op_descs = prog.Block(0).AllOps();
ir::Node* found_node = nullptr;
for (auto node : nodes) {
if (node->IsOp() && node->outputs.back()->Name() == "e") {
found_node = node;
break;
}
}
PADDLE_ENFORCE(found_node != nullptr);
for (auto it = op_descs.begin(); it != op_descs.end();) {
if (IsSameDesc(*it, found_node->Op())) {
it = op_descs.erase(it);
} else {
++it;
}
}
auto find_node_in_graph = [&](std::string s) {
ir::Node* ret = nullptr;
for (auto n : graph.Nodes()) {
if (n->Name() == s) {
ret = n;
break;
}
}
PADDLE_ENFORCE(ret != nullptr);
return ret;
};
ir::Node* e = find_node_in_graph("e");
ir::Node* d = find_node_in_graph("d");
std::remove(d->outputs.begin(), d->outputs.end(), found_node);
graph.RemoveNode(found_node);
graph.RemoveNode(e);
// other node keeps the same order
auto remain_nodes = SortOpLikeDescOrder(graph);
for (size_t i = 0; i < remain_nodes.size(); ++i) {
auto node = remain_nodes[i];
auto op_desc = op_descs[i];
ASSERT_TRUE(IsSameDesc(node->Op(), op_desc));
}
}
// 3. add some op_desc
TEST(SortOpLikeDescOrder, AddOpDesc) {
auto prog = FillProgramDesc();
const std::vector<OpDesc*>* all_op_descs =
new std::vector<OpDesc*>(prog.Block(0).AllOps());
ir::Graph graph(prog);
auto find_node_in_graph = [&](std::string s) {
ir::Node* ret = nullptr;
for (auto n : graph.Nodes()) {
if (n->Name() == s) {
ret = n;
break;
}
}
PADDLE_ENFORCE(ret != nullptr);
return ret;
};
// cached desc different with real one
// mimic the intermidiete pass modify the programdesc.
graph.Set(details::kAllOpDescs, all_op_descs); // take ownership
auto op_descs = prog.Block(0).AllOps();
auto op = prog.MutableBlock(0)->AppendOp();
prog.MutableBlock(0)->Var("d1")->SetType(proto::VarType::LOD_TENSOR);
op->SetType("sum");
op->SetInput("X", {"b", "c"});
op->SetOutput("Out", {"d1"});
ir::Node* node = graph.CreateOpNode(op);
ir::Node* d1 = graph.CreateVarNode(prog.MutableBlock(0)->Var("d1"));
ir::Node* b = find_node_in_graph("b");
ir::Node* c = find_node_in_graph("c");
node->outputs.emplace_back(d1);
node->inputs.emplace_back(b);
node->inputs.emplace_back(c);
d1->inputs.emplace_back(node);
b->outputs.emplace_back(node);
c->outputs.emplace_back(node);
op_descs.insert(op_descs.begin() + 4, op);
auto nodes = SortOpLikeDescOrder(graph);
for (size_t i = 0; i < nodes.size(); ++i) {
auto node = nodes[i];
auto op_desc = op_descs[i];
ASSERT_TRUE(IsSameDesc(node->Op(), op_desc));
}
}
// 4. add and delete some op_desc
TEST(SortOpLikeDescOrder, AddAndDeleteOpDesc) {
auto prog = FillProgramDesc();
ir::Graph graph(prog);
const std::vector<OpDesc*>* all_op_descs =
new std::vector<OpDesc*>(prog.Block(0).AllOps());
graph.Set(details::kAllOpDescs, all_op_descs); // take ownership
auto find_node_in_graph = [&](std::string s) {
ir::Node* ret = nullptr;
for (auto n : graph.Nodes()) {
if (n->Name() == s) {
ret = n;
break;
}
}
PADDLE_ENFORCE(ret != nullptr);
return ret;
};
// remove sum node
auto op_descs = prog.Block(0).AllOps();
ir::Node* found_node = nullptr;
auto nodes = graph.Nodes();
for (auto node : nodes) {
if (node->Name() == "sum") {
found_node = node;
break;
}
}
PADDLE_ENFORCE(found_node != nullptr);
for (auto it = op_descs.begin(); it != op_descs.end();) {
if (IsSameDesc(*it, found_node->Op())) {
it = op_descs.erase(it);
} else {
++it;
}
}
{
ir::Node* d = find_node_in_graph("d");
ir::Node* c = find_node_in_graph("c");
ir::Node* e = find_node_in_graph("e");
std::remove(d->outputs.begin(), d->outputs.end(), found_node);
std::remove(c->outputs.begin(), c->outputs.end(), found_node);
ir::Node* pending_op = found_node->outputs[0]->outputs[0];
graph.RemoveNode(e);
graph.RemoveNode(pending_op);
graph.RemoveNode(found_node);
}
// add node
auto op = prog.MutableBlock(0)->AppendOp();
prog.MutableBlock(0)->Var("d1")->SetType(proto::VarType::LOD_TENSOR);
op->SetType("sum");
op->SetInput("X", {"b", "c"});
op->SetOutput("Out", {"d1"});
{
ir::Node* node = graph.CreateOpNode(op);
ir::Node* d1 = graph.CreateVarNode(prog.MutableBlock(0)->Var("d1"));
ir::Node* b = find_node_in_graph("b");
ir::Node* c = find_node_in_graph("c");
node->outputs.emplace_back(d1);
node->inputs.emplace_back(b);
node->inputs.emplace_back(c);
b->outputs.emplace_back(node);
c->outputs.emplace_back(node);
}
op_descs.insert(op_descs.begin() + 2, op);
// check the order
auto mynodes = SortOpLikeDescOrder(graph);
for (size_t i = 0; i < mynodes.size(); ++i) {
auto node = mynodes[i];
auto op_desc = op_descs[i];
ASSERT_TRUE(IsSameDesc(node->Op(), op_desc));
}
}
// 5. add and replace some op_desc inplace.
TEST(SortOpLikeDescOrder, AddAndReplaceOpDescInplace) {
auto prog = FillProgramDesc();
ir::Graph graph(prog);
const std::vector<OpDesc*>* all_op_descs =
new std::vector<OpDesc*>(prog.Block(0).AllOps());
graph.Set(details::kAllOpDescs, all_op_descs); // take ownership
auto find_node_in_graph = [&](std::string s) {
ir::Node* ret = nullptr;
for (auto n : graph.Nodes()) {
if (n->Name() == s) {
ret = n;
break;
}
}
PADDLE_ENFORCE(ret != nullptr);
return ret;
};
auto op_descs = prog.Block(0).AllOps();
// add node
auto op = prog.MutableBlock(0)->AppendOp();
prog.MutableBlock(0)->Var("d1")->SetType(proto::VarType::LOD_TENSOR);
op->SetType("sum");
op->SetInput("X", {"b", "c"});
op->SetOutput("Out", {"d1"});
{
ir::Node* node = graph.CreateOpNode(op);
ir::Node* d1 = graph.CreateVarNode(prog.MutableBlock(0)->Var("d1"));
ir::Node* b = find_node_in_graph("b");
ir::Node* c = find_node_in_graph("c");
node->outputs.emplace_back(d1);
node->inputs.emplace_back(b);
node->inputs.emplace_back(c);
d1->inputs.emplace_back(node);
b->outputs.emplace_back(node);
c->outputs.emplace_back(node);
}
op_descs.emplace_back(op);
// replace op_desc inplace
auto nodes = graph.Nodes();
ir::Node* found_node = nullptr;
for (auto node : nodes) {
if (node->IsOp() && node->Op() && node->Name() == "assign") {
if (node->outputs.size() == 1 && node->outputs[0]->Name() == "e") {
found_node = node;
break;
}
}
}
{
ir::Node* d = find_node_in_graph("d");
ir::Node* e = find_node_in_graph("e");
std::remove(d->outputs.begin(), d->outputs.end(), found_node);
std::remove(e->inputs.begin(), e->inputs.end(), found_node);
graph.RemoveNode(found_node);
}
op_descs.erase(op_descs.begin() + 3);
auto replace_op = prog.MutableBlock(0)->AppendOp();
replace_op->SetType("sum");
replace_op->SetInput("X", {"d", "d1"});
replace_op->SetOutput("Out", {"e"});
{
ir::Node* sum2 = graph.CreateOpNode(replace_op);
ir::Node* e = find_node_in_graph("e");
ir::Node* d = find_node_in_graph("d");
ir::Node* d1 = find_node_in_graph("d1");
sum2->inputs.emplace_back(d);
sum2->inputs.emplace_back(d1);
sum2->outputs.emplace_back(e);
e->inputs.emplace_back(sum2);
d->outputs.emplace_back(sum2);
d1->outputs.emplace_back(sum2);
}
op_descs.emplace_back(replace_op);
// compare op order
auto graph_nodes = SortOpLikeDescOrder(graph);
for (size_t i = 0; i < graph_nodes.size(); ++i) {
auto node = graph_nodes[i];
auto op_desc = op_descs[i];
ASSERT_TRUE(IsSameDesc(node->Op(), op_desc));
}
}
} // namespace details
} // namespace framework
} // namespace paddle
...@@ -14,11 +14,16 @@ limitations under the License. */ ...@@ -14,11 +14,16 @@ limitations under the License. */
#include "paddle/fluid/framework/details/build_strategy.h" #include "paddle/fluid/framework/details/build_strategy.h"
#include <glog/logging.h>
#include <memory>
#include "paddle/fluid/framework/details/memory_reuse_types.h"
#include "paddle/fluid/framework/details/multi_devices_graph_check_pass.h" #include "paddle/fluid/framework/details/multi_devices_graph_check_pass.h"
#include "paddle/fluid/framework/details/multi_devices_graph_print_pass.h" #include "paddle/fluid/framework/details/multi_devices_graph_print_pass.h"
#include "paddle/fluid/framework/details/reduce_op_handle.h" #include "paddle/fluid/framework/details/reduce_op_handle.h"
#include "paddle/fluid/framework/details/sequential_execution_pass.h" #include "paddle/fluid/framework/details/sequential_execution_pass.h"
#include "paddle/fluid/framework/ir/graph.h" #include "paddle/fluid/framework/ir/graph.h"
#include "paddle/fluid/framework/ir/graph_helper.h"
#include "paddle/fluid/framework/ir/graph_viz_pass.h" #include "paddle/fluid/framework/ir/graph_viz_pass.h"
namespace paddle { namespace paddle {
...@@ -69,6 +74,14 @@ class ParallelExecutorPassBuilder : public ir::PassBuilder { ...@@ -69,6 +74,14 @@ class ParallelExecutorPassBuilder : public ir::PassBuilder {
} }
VLOG(1) << "CollectiveContext:" << context->String(); VLOG(1) << "CollectiveContext:" << context->String();
// NOTE(dzh): memory optimize should be a runtime pass.
// However, after multi_devices_pass, VarHandle, OpHandle is
// the de-fact IR, any reuse on Graph is meaningless.
// A side-effect of that, memory optimize cannot forsee the fetched vars
// , so fetchlist should be set persistable before call the Run interface.
if (strategy.memory_optimize_) {
auto analysis_var_pass = AppendPass("analysis_var_pass");
}
// Convert graph to run on multi-devices. // Convert graph to run on multi-devices.
auto multi_devices_pass = AppendPass("multi_devices_pass"); auto multi_devices_pass = AppendPass("multi_devices_pass");
multi_devices_pass->SetNotOwned<const BuildStrategy>("strategy", multi_devices_pass->SetNotOwned<const BuildStrategy>("strategy",
...@@ -79,8 +92,11 @@ class ParallelExecutorPassBuilder : public ir::PassBuilder { ...@@ -79,8 +92,11 @@ class ParallelExecutorPassBuilder : public ir::PassBuilder {
// Add a graph print pass to record a graph with device info. // Add a graph print pass to record a graph with device info.
if (!strategy_.debug_graphviz_path_.empty()) { if (!strategy_.debug_graphviz_path_.empty()) {
auto multi_devices_print_pass = AppendPass("multi_devices_print_pass"); auto multi_devices_print_pass = AppendPass("multi_devices_print_pass");
multi_devices_print_pass->SetNotOwned<const std::string>( const std::string graph_path =
"debug_graphviz_path", &strategy_.debug_graphviz_path_); string::Sprintf("%s%s", strategy_.debug_graphviz_path_.c_str(),
"_multi_devices_graph");
multi_devices_print_pass->Set<std::string>(kGraphvizPath,
new std::string(graph_path));
multi_devices_print_pass->Set<details::GraphvizSSAGraphPrinter>( multi_devices_print_pass->Set<details::GraphvizSSAGraphPrinter>(
"graph_printer", new details::GraphvizSSAGraphPrinter); "graph_printer", new details::GraphvizSSAGraphPrinter);
} }
...@@ -127,7 +143,6 @@ std::unique_ptr<ir::Graph> BuildStrategy::Apply( ...@@ -127,7 +143,6 @@ std::unique_ptr<ir::Graph> BuildStrategy::Apply(
CreatePassesFromStrategy(false); CreatePassesFromStrategy(false);
std::unique_ptr<ir::Graph> graph(new ir::Graph(main_program)); std::unique_ptr<ir::Graph> graph(new ir::Graph(main_program));
for (std::shared_ptr<ir::Pass> &pass : pass_builder_->AllPasses()) { for (std::shared_ptr<ir::Pass> &pass : pass_builder_->AllPasses()) {
if (pass->Type() == "multi_devices_pass") { if (pass->Type() == "multi_devices_pass") {
pass->Erase("places"); pass->Erase("places");
...@@ -145,6 +160,17 @@ std::unique_ptr<ir::Graph> BuildStrategy::Apply( ...@@ -145,6 +160,17 @@ std::unique_ptr<ir::Graph> BuildStrategy::Apply(
pass->Erase("nccl_ctxs"); pass->Erase("nccl_ctxs");
pass->SetNotOwned<platform::NCCLContextMap>("nccl_ctxs", nctx); pass->SetNotOwned<platform::NCCLContextMap>("nccl_ctxs", nctx);
#endif #endif
} else if (pass->Type() == "analysis_var_pass") {
const std::vector<OpDesc *> *all_op_descs =
new std::vector<OpDesc *>(main_program.Block(0).AllOps());
graph->Set<const std::vector<OpDesc *>>(kAllOpDescs,
all_op_descs); // take ownership
graph->Set<GraphNodePool>(kGraphNodePool,
new GraphNodePool); // take ownership
pass->Erase(kAllOpDescs);
pass->SetNotOwned<const std::vector<OpDesc *>>(kAllOpDescs, all_op_descs);
} else if (pass->Type() == "sequential_execution_pass") { } else if (pass->Type() == "sequential_execution_pass") {
LOG(INFO) << "set enable_sequential_execution:" LOG(INFO) << "set enable_sequential_execution:"
<< enable_sequential_execution_; << enable_sequential_execution_;
...@@ -166,6 +192,7 @@ std::unique_ptr<ir::Graph> BuildStrategy::Apply( ...@@ -166,6 +192,7 @@ std::unique_ptr<ir::Graph> BuildStrategy::Apply(
} }
return graph; return graph;
} }
} // namespace details } // namespace details
} // namespace framework } // namespace framework
} // namespace paddle } // namespace paddle
...@@ -176,6 +203,7 @@ USE_PASS(multi_batch_merge_pass); ...@@ -176,6 +203,7 @@ USE_PASS(multi_batch_merge_pass);
USE_PASS(multi_devices_pass); USE_PASS(multi_devices_pass);
USE_PASS(multi_devices_check_pass); USE_PASS(multi_devices_check_pass);
USE_PASS(multi_devices_print_pass); USE_PASS(multi_devices_print_pass);
USE_PASS(analysis_var_pass);
USE_PASS(sequential_execution_pass); USE_PASS(sequential_execution_pass);
USE_PASS(all_reduce_deps_pass); USE_PASS(all_reduce_deps_pass);
USE_PASS(modify_op_lock_and_record_event_pass); USE_PASS(modify_op_lock_and_record_event_pass);
...@@ -60,8 +60,15 @@ struct BuildStrategy { ...@@ -60,8 +60,15 @@ struct BuildStrategy {
kCustomized = 2, kCustomized = 2,
}; };
enum class OptimizeStrategy {
// To be Implemented,bruteforce, recursive compute unused var names.
kBruteForce = 0,
kControlFlowGraph = 1, // use cfg_graph algorithm, faster speed.
};
ReduceStrategy reduce_{ReduceStrategy::kAllReduce}; ReduceStrategy reduce_{ReduceStrategy::kAllReduce};
GradientScaleStrategy gradient_scale_{GradientScaleStrategy::kCoeffNumDevice}; GradientScaleStrategy gradient_scale_{GradientScaleStrategy::kCoeffNumDevice};
OptimizeStrategy strategy_{OptimizeStrategy::kControlFlowGraph};
std::string debug_graphviz_path_{""}; std::string debug_graphviz_path_{""};
...@@ -69,6 +76,10 @@ struct BuildStrategy { ...@@ -69,6 +76,10 @@ struct BuildStrategy {
bool enable_data_balance_{false}; bool enable_data_balance_{false};
bool memory_optimize_{false};
bool memory_early_delete_{false};
bool enable_sequential_execution_{false}; bool enable_sequential_execution_{false};
bool fuse_broadcast_op_{false}; bool fuse_broadcast_op_{false};
......
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#pragma once
#include <string>
#include <vector>
#include "paddle/fluid/framework/details/computation_op_handle.h"
#include "paddle/fluid/framework/details/op_handle_base.h"
#include "paddle/fluid/framework/details/var_handle.h"
#include "paddle/fluid/framework/garbage_collector.h"
#include "paddle/fluid/framework/lod_tensor_array.h"
#include "paddle/fluid/framework/scope.h"
#include "paddle/fluid/framework/selected_rows.h"
#include "paddle/fluid/framework/tensor.h"
namespace paddle {
namespace framework {
namespace details {
class EarlyDeleteOpHandle : public OpHandleBase {
public:
EarlyDeleteOpHandle(ir::Node* node, const Scope* scope,
const platform::Place& place,
const std::vector<std::string>& names,
GarbageCollector* gc)
: OpHandleBase(node),
scope_(scope),
place_(place),
names_(names),
gc_(gc) {
#ifdef PADDLE_WITH_CUDA
if (IsStreamGarabageCollector()) {
auto gpu_place = boost::get<platform::CUDAPlace>(place);
PADDLE_ENFORCE(cudaSetDevice(gpu_place.device));
PADDLE_ENFORCE(cudaEventCreateWithFlags(&event_, cudaEventDisableTiming));
}
#endif
}
~EarlyDeleteOpHandle() {
#ifdef PADDLE_WITH_CUDA
if (IsStreamGarabageCollector()) {
auto gpu_place = boost::get<platform::CUDAPlace>(dev_ctx_->GetPlace());
PADDLE_ENFORCE(cudaSetDevice(gpu_place.device));
PADDLE_ENFORCE(cudaEventDestroy(event_));
}
#endif
}
std::string Name() const override { return "early_delete"; }
protected:
void RunImpl() override {
std::vector<std::shared_ptr<memory::Allocation>> tensors;
auto* local_scope = scope_->FindVar(kLocalExecScopeName)->Get<Scope*>();
for (auto& var_name : names_) {
auto* var = local_scope->FindVar(var_name);
PADDLE_ENFORCE(var != nullptr,
string::Sprintf("Local Scope not has var %s", var_name));
if (var->IsType<LoDTensor>()) {
tensors.emplace_back(var->GetMutable<LoDTensor>()->MoveMemoryHolder());
} else if (var->IsType<SelectedRows>()) {
tensors.emplace_back(var->GetMutable<SelectedRows>()
->mutable_value()
->MoveMemoryHolder());
} else if (var->IsType<LoDTensorArray>()) {
LoDTensorArray* tensor_array = var->GetMutable<LoDTensorArray>();
for (auto& tensor : *tensor_array) {
tensors.emplace_back(tensor.MoveMemoryHolder());
}
}
}
if (!tensors.empty()) {
ClearTensors(tensors);
}
}
private:
void ClearTensors(
const std::vector<std::shared_ptr<memory::Allocation>>& tensors) {
if (platform::is_cpu_place(place_)) {
ClearCPUTensors(tensors);
} else {
ClearGPUTensors(tensors);
}
}
void ClearCPUTensors(
const std::vector<std::shared_ptr<memory::Allocation>>& tensors) {
auto* gc = dynamic_cast<CPUGarbageCollector*>(gc_);
if (gc != nullptr) {
gc->Add(tensors);
}
}
void ClearGPUTensors(
const std::vector<std::shared_ptr<memory::Allocation>>& tensors) {
#ifdef PADDLE_WITH_CUDA
auto* gc = dynamic_cast<StreamGarbageCollector*>(gc_);
if (gc != nullptr) {
auto compute_stream = dev_ctx_->stream();
auto callback_stream = gc->stream();
auto callback_func = [=]() {
PADDLE_ENFORCE(cudaEventRecord(event_, compute_stream));
PADDLE_ENFORCE(cudaStreamWaitEvent(callback_stream, event_, 0));
};
gc_->Add(tensors, callback_func);
} else {
gc_->Add(tensors);
}
}
bool IsStreamGarabageCollector() const {
return dynamic_cast<const StreamGarbageCollector*>(gc_) != nullptr;
#endif
}
const Scope* scope_;
const platform::Place place_;
std::vector<std::string> names_;
GarbageCollector* gc_;
#ifdef PADDLE_WITH_CUDA
platform::CUDADeviceContext* dev_ctx_;
cudaEvent_t event_;
#endif
};
} // namespace details
} // namespace framework
} // namespace paddle
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#include "paddle/fluid/framework/details/memory_early_delete_pass.h"
#include <queue>
#include <string>
#include <vector>
#include "paddle/fluid/framework/details/memory_reuse_types.h"
#include "paddle/fluid/framework/details/multi_devices_helper.h"
#include "paddle/fluid/framework/details/reference_count_pass_helper.h"
#include "paddle/fluid/framework/ir/graph_helper.h"
namespace paddle {
namespace framework {
namespace details {
static ComputationOpHandle* FindNextComputationOpHandle(VarHandle* var_in) {
std::queue<VarHandleBase*> queue;
queue.push(var_in);
do {
auto* var = queue.front();
queue.pop();
for (auto* op : var->PendingOps()) {
auto* compute_op = dynamic_cast<ComputationOpHandle*>(op);
if (compute_op != nullptr && compute_op->GetPlace() == var_in->place_) {
return compute_op;
}
for (auto* out_var : op->Outputs()) {
queue.push(out_var);
}
}
} while (!queue.empty());
return nullptr;
}
std::unique_ptr<ir::Graph> MemoryEarlyDeletePass::ApplyImpl(
std::unique_ptr<ir::Graph> graph) const {
auto& graph_pool = Get<GraphNodePool>(kGraphNodePool);
auto& gcs = Get<GarbageCollectorMap>(kGarbageCollector);
std::unordered_map<std::string, std::unordered_set<OpDesc*>> unlived_vars;
unlived_vars.reserve(graph_pool.size());
for (auto& pair : graph_pool) {
unlived_vars.insert(std::make_pair(pair.first, pair.second));
}
auto compare_and_insert_early_delete_op = [&](
OpHandleBase* op, const std::vector<VarHandleBase*>& vars) {
if (unlived_vars.empty()) return;
// unlived vars can be deleted after the last used op has finished.
auto* compute_op = dynamic_cast<ComputationOpHandle*>(op);
const auto& places = Get<std::vector<platform::Place>>(kAllPlaces);
for (auto& var : vars) {
auto* var_handle = dynamic_cast<VarHandle*>(var);
auto var_name = var->Node()->Name();
auto& var_place = var_handle->place_;
if (unlived_vars.count(var_name) == 0) continue;
if (!unlived_vars[var_name].empty()) {
if (compute_op != nullptr &&
unlived_vars[var_name].count(compute_op->Node()->Op()) != 0) {
unlived_vars[var_name].erase(compute_op->Node()->Op());
}
continue;
}
if (var_handle == nullptr || !var_handle->Node()->IsVar() ||
var_handle->Node()->IsCtrlVar())
continue;
// shameless copyed from reference count pass.
if (compute_op == nullptr) {
// use next computation op scope
compute_op = FindNextComputationOpHandle(var_handle);
}
auto* early_delete_node =
graph->CreateEmptyNode("early_delete", ir::Node::Type::kOperation);
GarbageCollector* gc = gcs.at(places[compute_op->GetScopeIdx()]).get();
auto* early_delete_handle = new EarlyDeleteOpHandle(
early_delete_node, compute_op->GetScope(), var_place, {var_name}, gc);
if (compute_op->Outputs().empty()) {
auto* dep_var = new DummyVarHandle(graph->CreateControlDepVar());
compute_op->AddOutput(dep_var);
graph->Get<GraphDepVars>(kGraphDepVars).emplace(dep_var);
}
early_delete_handle->AddInput(compute_op->Outputs().front());
VLOG(5) << "Add early delete op " << var_name << " to Operator"
<< compute_op->Name();
}
};
auto all_ops = ir::FilterByNodeWrapper<OpHandleBase>(*graph);
for (auto& op : all_ops) {
compare_and_insert_early_delete_op(op, op->Inputs());
compare_and_insert_early_delete_op(op, op->Outputs());
}
return graph;
}
} // namespace details
} // namespace framework
} // namespace paddle
REGISTER_PASS(memory_early_delete_pass,
paddle::framework::details::MemoryEarlyDeletePass)
.RequireGraphAttr(paddle::framework::details::kGraphNodePool)
.RequireGraphAttr(paddle::framework::details::kGarbageCollector);
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#pragma once
#include "paddle/fluid/framework/details/early_delete_op_handle.h"
#include "paddle/fluid/framework/ir/graph.h"
#include "paddle/fluid/framework/ir/pass.h"
namespace paddle {
namespace framework {
namespace details {
class MemoryEarlyDeletePass : public ir::Pass {
protected:
std::unique_ptr<ir::Graph> ApplyImpl(
std::unique_ptr<ir::Graph> graph) const override;
};
} // namespace details
} // namespace framework
} // namespace paddle
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#include "paddle/fluid/framework/details/memory_reuse_types.h"
#include <iostream>
#include <sstream>
#include <string>
namespace paddle {
namespace framework {
namespace details {
size_t NodeSizeInBytes(ir::Node* n) {
auto* desc = FindVarDescInBlock(n);
auto shape = desc->GetShape();
size_t type_size = SizeOfType(desc->GetDataType());
int size = 1;
for (auto& s : shape) {
size *= s;
}
return type_size * std::abs(size);
}
std::string DebugStringImpl(VarDesc* var) {
std::stringstream ss;
ss << var->Name();
ss << "[";
try {
auto shape = var->GetShape();
for (size_t i = 0; i < shape.size(); ++i) {
if (i != shape.size() - 1) {
ss << shape[i] << ",";
} else {
ss << shape[i];
}
}
ss << "]";
} catch (...) {
ss << "Var has no VarDesc !!! Name:" << var->Name();
}
return ss.str();
}
std::string DebugString(ir::Node* var) {
return DebugStringImpl(FindVarDescInBlock(var));
}
// return DebugString(var->Var()); }
// NOTE(dzh): based ir node, if a large node has been reused
// by a small size node, then next time it appear in pool, it will
// have the small size. Find the original node shap from blockdesc.
VarDesc* FindVarDescInBlock(ir::Node* n) {
PADDLE_ENFORCE(n->IsVar() && !n->IsCtrlVar() && n->inputs.size() == 1);
BlockDesc* block = n->inputs[0]->Op()->Block();
PADDLE_ENFORCE(block->HasVar(n->Name()),
string::Sprintf("Block do not has var %s", n->Name()));
return block->FindVar(n->Name());
}
struct NodeComparator {
bool operator()(ir::Node* lhs, ir::Node* rhs) const {
auto* lhs_desc = FindVarDescInBlock(lhs);
auto* rhs_desc = FindVarDescInBlock(rhs);
auto lhs_shape = lhs_desc->GetShape();
auto rhs_shape = rhs_desc->GetShape();
if ((lhs_shape[0] == -1 && rhs_shape[0] == -1) ||
(lhs_shape[0] != -1 && rhs_shape[0] != -1)) {
return NodeSizeInBytes(lhs) <= NodeSizeInBytes(rhs);
} else {
return false;
}
}
};
void OrderedNodePairPool::Insert(ir::Node* var, ir::Node* op) {
PADDLE_ENFORCE(var->IsVar() && !var->IsCtrlVar());
PADDLE_ENFORCE(op->IsOp());
if (mark_table_.count(var->Name()) != 0) {
mark_table_[var->Name()]->second.insert(op);
return;
}
auto* var_desc = FindVarDescInBlock(var);
auto var_shape = var_desc->GetShape();
int batch_size = static_cast<int>(var_shape[0]);
NodeComparator compare_node;
Iter it = nodes_.begin();
while (it != nodes_.end()) {
auto* cache_desc = FindVarDescInBlock(it->first);
int cache_batch_size = cache_desc->GetShape()[0];
if ((cache_batch_size == -1 && batch_size == -1) ||
(cache_batch_size != -1 && batch_size != -1)) {
if (compare_node(it->first, var)) {
++it;
} else {
break;
}
} else if (cache_batch_size == -1 && batch_size != -1) {
++it;
} else if (cache_batch_size != -1 && batch_size == -1) {
break;
}
}
it =
nodes_.insert(it, std::make_pair(var, std::unordered_set<ir::Node*>{op}));
mark_table_[var->Name()] = it;
}
int OrderedNodePairPool::GetIndex(ir::Node* var) {
return std::distance(nodes_.begin(), mark_table_[var->Name()]);
}
ir::Node* OrderedNodePairPool::NodeMatch(ir::Node* var) const {
ir::Node* found_node = nullptr;
NodeComparator compare_node;
for (auto it = nodes_.begin(); it != nodes_.end(); ++it) {
if (compare_node(var, it->first)) {
found_node = it->first;
break;
}
}
return found_node;
}
void OrderedNodePairPool::Erase(ir::Node* var) {
PADDLE_ENFORCE(mark_table_.count(var->Name()));
nodes_.erase(mark_table_[var->Name()]);
mark_table_.erase(var->Name());
}
std::string OrderedNodePairPool::ToString() const {
std::stringstream ss;
for (auto it = nodes_.begin(); it != nodes_.end(); ++it) {
ss << DebugString(it->first) << " ";
}
return ss.str();
}
} // namespace details
} // namespace framework
} // namespace paddle
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#pragma once
#include <algorithm>
#include <iostream>
#include <iterator>
#include <list>
#include <string>
#include <utility>
#include <vector>
#include "paddle/fluid/framework/data_type.h"
#include "paddle/fluid/framework/ir/graph.h"
namespace paddle {
namespace framework {
namespace details {
constexpr char kFetchedVars[] = "fetched_vars";
constexpr char kGraphNodePool[] = "graph_node_pool";
// NOTE(dzh): Variable and the operators use the var.
// for early delete pass.
// Because analysis var pass build base on ir::Node, which maybe released
// or modified between passes, so we use OpDesc* to mark ops.
using GraphNodePool = std::vector<
std::pair<std::string /*var node*/, std::unordered_set<OpDesc*> /* ops */>>;
// NOTE(dzh): by default, it sort node in ascend order(by node bytes size).
// in fluid, -1 means the batch_size is determined in runtime.
// the node batch_size equal -1 always ranking in the front than the node not.
// For example,
// node0[-1, 1] node1[-1, 1, 1], node2[1,1], node3[1,1024], ..
// O(1) insert, delete
class OrderedNodePairPool {
public:
using NodePair = std::pair<ir::Node*, std::unordered_set<ir::Node*>>;
using Iter = typename std::list<NodePair>::iterator;
using ConstIter = typename std::list<NodePair>::const_iterator;
void Insert(ir::Node* var, ir::Node* op);
void Erase(ir::Node* var);
bool Has(ir::Node* var) { return mark_table_.count(var->Name()); }
ir::Node* NodeMatch(ir::Node* var) const;
// map store non-const iterator, can not promise const
int GetIndex(ir::Node* var);
// pool all node to string
std::string ToString() const;
Iter begin() { return nodes_.begin(); }
Iter end() { return nodes_.end(); }
ConstIter begin() const { return nodes_.begin(); }
ConstIter end() const { return nodes_.end(); }
size_t size() const { return nodes_.size(); }
private:
// for searching.
std::unordered_map<std::string, Iter> mark_table_;
// node swap pairs. var -> ops dep var
std::list<NodePair> nodes_;
};
// node memory size in bytes
size_t NodeSizeInBytes(ir::Node* n);
std::string DebugString(ir::Node* var);
// std::string DebugString(VarDesc* var);
VarDesc* FindVarDescInBlock(ir::Node* n);
} // namespace details
} // namespace framework
} // namespace paddle
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#include "paddle/fluid/framework/details/memory_reuse_types.h"
#include <algorithm>
#include <iostream>
#include <memory>
#include <sstream>
#include <string>
#include <utility>
#include <vector>
#include "glog/logging.h"
#include "gtest/gtest.h"
namespace paddle {
namespace framework {
namespace details {
TEST(OrderedNodePairPool, Normal) {
OrderedNodePairPool pool;
std::vector<std::unique_ptr<ir::Node>> nodes;
// clang-format off
std::vector<std::vector<int64_t>> shapes = {{-1, 10},
{-1, 20},
{1, 2},
{5, 2},
{10, 20},
{-1, 2, 5},
{-1, 1, 5},
{-1, 1}};
// clang-format on
const int COUNT = shapes.size();
ProgramDesc prog;
BlockDesc* block_desc = prog.MutableBlock(0);
auto* op_desc = block_desc->AppendOp();
op_desc->SetType("dummy");
std::unique_ptr<ir::Node> op = ir::CreateNodeForTest(op_desc);
for (int i = 0; i < COUNT; ++i) {
auto desc = block_desc->Var(std::to_string(i));
desc->SetShape(shapes[i]);
std::unique_ptr<ir::Node> node = ir::CreateNodeForTest(desc);
node->inputs.emplace_back(op.get());
nodes.emplace_back(std::move(node));
}
for (auto& node : nodes) {
pool.Insert(node.get(), op.get());
}
// assert its order and interface.
std::cout << pool.ToString() << std::endl;
pool.Erase(nodes.front().get());
std::cout << pool.ToString() << std::endl;
ASSERT_EQ(pool.size(), static_cast<size_t>(COUNT - 1));
ASSERT_EQ(pool.GetIndex(nodes.back().get()), 0);
{
auto v1 = block_desc->Var("11");
v1->SetShape({-1, 256, 56, 56});
std::unique_ptr<ir::Node> node1 = ir::CreateNodeForTest(v1);
node1->inputs.emplace_back(op.get());
auto* cache = pool.NodeMatch(node1.get());
ASSERT_EQ(cache, nullptr);
}
{
auto v2 = block_desc->Var("12");
v2->SetShape({-1, 2, 5});
std::unique_ptr<ir::Node> node1 = ir::CreateNodeForTest(v2);
node1->inputs.emplace_back(op.get());
auto* cache = pool.NodeMatch(node1.get());
ASSERT_EQ(pool.GetIndex(cache), 2); // match 6:[-1,2,5]
}
{
auto v3 = block_desc->Var("13");
v3->SetShape({2, 5});
std::unique_ptr<ir::Node> node1 = ir::CreateNodeForTest(v3);
node1->inputs.emplace_back(op.get());
auto* cache = pool.NodeMatch(node1.get());
ASSERT_EQ(pool.GetIndex(cache), 5); // match 4:[5,2]
}
}
} // namespace details
} // namespace framework
} // namespace paddle
...@@ -85,4 +85,5 @@ void GraphvizSSAGraphPrinter::Print(const ir::Graph &graph, ...@@ -85,4 +85,5 @@ void GraphvizSSAGraphPrinter::Print(const ir::Graph &graph,
} // namespace paddle } // namespace paddle
REGISTER_PASS(multi_devices_print_pass, REGISTER_PASS(multi_devices_print_pass,
paddle::framework::details::SSAGraghBuilderWithPrinter); paddle::framework::details::SSAGraghBuilderWithPrinter)
.RequirePassAttr(paddle::framework::details::kGraphvizPath);
...@@ -14,6 +14,7 @@ ...@@ -14,6 +14,7 @@
#pragma once #pragma once
#include <glog/logging.h>
#include <fstream> #include <fstream>
#include <iosfwd> #include <iosfwd>
#include <ostream> #include <ostream>
...@@ -24,6 +25,8 @@ namespace paddle { ...@@ -24,6 +25,8 @@ namespace paddle {
namespace framework { namespace framework {
namespace details { namespace details {
constexpr char kGraphvizPath[] = "debug_graphviz_path";
class SSAGraphPrinter { class SSAGraphPrinter {
public: public:
virtual ~SSAGraphPrinter() {} virtual ~SSAGraphPrinter() {}
...@@ -40,7 +43,7 @@ class SSAGraghBuilderWithPrinter : public ir::Pass { ...@@ -40,7 +43,7 @@ class SSAGraghBuilderWithPrinter : public ir::Pass {
std::unique_ptr<ir::Graph> ApplyImpl( std::unique_ptr<ir::Graph> ApplyImpl(
std::unique_ptr<ir::Graph> graph) const override { std::unique_ptr<ir::Graph> graph) const override {
std::unique_ptr<std::ostream> fout( std::unique_ptr<std::ostream> fout(
new std::ofstream(Get<const std::string>("debug_graphviz_path"))); new std::ofstream(Get<std::string>(kGraphvizPath)));
PADDLE_ENFORCE(fout->good()); PADDLE_ENFORCE(fout->good());
Get<GraphvizSSAGraphPrinter>("graph_printer").Print(*graph, *fout); Get<GraphvizSSAGraphPrinter>("graph_printer").Print(*graph, *fout);
return graph; return graph;
......
...@@ -25,7 +25,7 @@ namespace paddle { ...@@ -25,7 +25,7 @@ namespace paddle {
namespace framework { namespace framework {
namespace details { namespace details {
constexpr char kLocalExecScopeName[] = "@LCOAL_SCOPE@"; constexpr char kLocalExecScopeName[] = "@LOCAL_SCOPE@";
// Wraps ir::Node and provide helper utilities. // Wraps ir::Node and provide helper utilities.
// It's responsible for populating necessary fields of ir::Node. // It's responsible for populating necessary fields of ir::Node.
......
...@@ -100,7 +100,7 @@ static void DeleteUnusedTensors( ...@@ -100,7 +100,7 @@ static void DeleteUnusedTensors(
continue; continue;
} }
auto* var = scope.FindVar(name); auto* var = scope.FindVar(name);
if (var != nullptr) { if (var == nullptr) {
continue; continue;
} }
......
...@@ -44,6 +44,7 @@ pass_library(seqconv_eltadd_relu_fuse_pass inference) ...@@ -44,6 +44,7 @@ pass_library(seqconv_eltadd_relu_fuse_pass inference)
pass_library(is_test_pass base) pass_library(is_test_pass base)
pass_library(conv_elementwise_add_act_fuse_pass inference) pass_library(conv_elementwise_add_act_fuse_pass inference)
pass_library(conv_elementwise_add2_act_fuse_pass inference) pass_library(conv_elementwise_add2_act_fuse_pass inference)
pass_library(conv_elementwise_add_fuse_pass inference)
if(WITH_MKLDNN) if(WITH_MKLDNN)
pass_library(mkldnn_placement_pass base) pass_library(mkldnn_placement_pass base)
pass_library(depthwise_conv_mkldnn_pass base) pass_library(depthwise_conv_mkldnn_pass base)
......
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#include <string>
#include "paddle/fluid/framework/ir/conv_elementwise_add_fuse_pass.h"
#include "paddle/fluid/framework/ir/graph_viz_pass.h"
namespace paddle {
namespace framework {
namespace ir {
#define GET_IR_NODE(node__) GET_IR_NODE_FROM_SUBGRAPH(node__, node__, pattern);
#define GET_NODES \
GET_IR_NODE(conv_op); \
GET_IR_NODE(conv_out); \
GET_IR_NODE(conv_filter); \
GET_IR_NODE(elementwise_add_op); \
GET_IR_NODE(elementwise_add_in_y); \
GET_IR_NODE(elementwise_add_out);
std::unique_ptr<ir::Graph> ConvElementwiseAddFusePass::ApplyImpl(
std::unique_ptr<ir::Graph> graph) const {
const std::string pattern_name = "conv_elementwise_add_fuse";
FusePassBase::Init(pattern_name, graph.get());
GraphPatternDetector gpd;
auto* x = gpd.mutable_pattern()
->NewNode("x")
->assert_is_op_input("conv2d", "Input")
->AsInput();
patterns::ConvElementwiseadd pattern(gpd.mutable_pattern(), pattern_name);
pattern(x);
auto handler = [&](const GraphPatternDetector::subgraph_t& subgraph,
Graph* g) {
GET_NODES;
auto base_op_desc = *conv_op->Op()->Proto();
std::string bias_name = elementwise_add_in_y->Name();
std::string output_name = elementwise_add_out->Name();
std::string act_type = "identity";
framework::OpDesc new_op_desc(base_op_desc, nullptr);
new_op_desc.SetType("conv2d_fusion");
new_op_desc.SetInput("Bias", {bias_name});
new_op_desc.SetInput("ResidualData", {});
new_op_desc.SetAttr("activation", act_type);
new_op_desc.SetOutput("Output", {output_name});
new_op_desc.SetAttr("is_test", true);
new_op_desc.SetAttr("use_cudnn", false);
new_op_desc.Flush();
// Create a new node for the fused op.
auto* new_conv_op = graph->CreateOpNode(&new_op_desc);
// Link inputs and outputs.
PADDLE_ENFORCE(subgraph.count(x));
auto* conv_in_node = subgraph.at(x);
IR_NODE_LINK_TO(conv_in_node, new_conv_op); // Input
IR_NODE_LINK_TO(conv_filter, new_conv_op); // Filter
IR_NODE_LINK_TO(elementwise_add_in_y, new_conv_op); // Bias
IR_NODE_LINK_TO(new_conv_op, elementwise_add_out); // Output
// Delete the unneeded nodes.
GraphSafeRemoveNodes(graph.get(), {conv_op, conv_out, elementwise_add_op});
};
gpd(graph.get(), handler);
return graph;
}
} // namespace ir
} // namespace framework
} // namespace paddle
REGISTER_PASS(conv_elementwise_add_fuse_pass,
paddle::framework::ir::ConvElementwiseAddFusePass);
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#pragma once
#include "paddle/fluid/framework/ir/fuse_pass_base.h"
#include "paddle/fluid/framework/ir/graph_pattern_detector.h"
namespace paddle {
namespace framework {
namespace ir {
class ConvElementwiseAddFusePass : public FusePassBase {
public:
virtual ~ConvElementwiseAddFusePass() {}
protected:
std::unique_ptr<ir::Graph> ApplyImpl(std::unique_ptr<ir::Graph> graph) const;
};
} // namespace ir
} // namespace framework
} // namespace paddle
...@@ -162,7 +162,10 @@ void Graph::ResolveHazard( ...@@ -162,7 +162,10 @@ void Graph::ResolveHazard(
(*it_new)->inputs.empty() ? nullptr : (*it_new)->inputs[0]; (*it_new)->inputs.empty() ? nullptr : (*it_new)->inputs[0];
const auto &read_ops = (*it_old)->outputs; const auto &read_ops = (*it_old)->outputs;
PADDLE_ENFORCE(write_op, "The write_op should not be empty."); PADDLE_ENFORCE(
write_op,
string::Sprintf("The write_op of var %s should not be empty.",
(*it_new)->Name()));
// Add write after write dependence // Add write after write dependence
ir::Node *upstream_op = ir::Node *upstream_op =
......
...@@ -18,6 +18,7 @@ limitations under the License. */ ...@@ -18,6 +18,7 @@ limitations under the License. */
#include <fstream> #include <fstream>
#include <iosfwd> #include <iosfwd>
#include <ostream> #include <ostream>
#include <unordered_map>
#include <unordered_set> #include <unordered_set>
DEFINE_string(print_sub_graph_dir, "", DEFINE_string(print_sub_graph_dir, "",
...@@ -121,7 +122,7 @@ std::map<ir::Node *, std::unordered_set<ir::Node *>> BuildOperationAdjList( ...@@ -121,7 +122,7 @@ std::map<ir::Node *, std::unordered_set<ir::Node *>> BuildOperationAdjList(
} }
size_t GraphNum(const Graph &graph) { size_t GraphNum(const Graph &graph) {
std::unordered_set<ir::Node *> nodes = graph.Nodes(); std::unordered_set<ir::Node *> nodes(graph.Nodes());
std::unordered_set<ir::Node *> visited_nodes; std::unordered_set<ir::Node *> visited_nodes;
visited_nodes.reserve(nodes.size()); visited_nodes.reserve(nodes.size());
std::deque<ir::Node *> q_nodes; std::deque<ir::Node *> q_nodes;
......
...@@ -24,6 +24,7 @@ limitations under the License. */ ...@@ -24,6 +24,7 @@ limitations under the License. */
namespace paddle { namespace paddle {
namespace framework { namespace framework {
namespace ir { namespace ir {
// Test if the graph contains circle. // Test if the graph contains circle.
bool HasCircle(const Graph &graph); bool HasCircle(const Graph &graph);
......
...@@ -17,7 +17,6 @@ ...@@ -17,7 +17,6 @@
#include <string> #include <string>
#include <vector> #include <vector>
#include "graph_pattern_detector.h"
#include "paddle/fluid/framework/ir/graph_helper.h" #include "paddle/fluid/framework/ir/graph_helper.h"
#include "paddle/fluid/framework/ir/graph_pattern_detector.h" #include "paddle/fluid/framework/ir/graph_pattern_detector.h"
#include "paddle/fluid/framework/ir/graph_traits.h" #include "paddle/fluid/framework/ir/graph_traits.h"
...@@ -1210,6 +1209,33 @@ PDNode *patterns::ConvElementwiseadd2Act::operator()(PDNode *conv_in) { ...@@ -1210,6 +1209,33 @@ PDNode *patterns::ConvElementwiseadd2Act::operator()(PDNode *conv_in) {
return act_out; return act_out;
} }
PDNode *patterns::ConvElementwiseadd::operator()(PDNode *conv_in) {
conv_in->AsInput();
auto conv_op = pattern->NewNode(conv_op_repr())->assert_is_op("conv2d");
auto conv_out = pattern->NewNode(conv_out_repr())
->assert_is_op_output("conv2d")
->assert_is_op_input("elementwise_add", "X")
->AsIntermediate();
auto conv_filter = pattern->NewNode(conv_filter_repr())
->assert_is_op_input("conv2d", "Filter")
->AsInput();
auto elementwise_add_op = pattern->NewNode(elementwise_add_op_repr())
->assert_is_op("elementwise_add");
auto elementwise_add_in_y = pattern->NewNode(elementwise_add_in_y_repr())
->assert_is_op_input("elementwise_add", "Y")
->AsInput();
auto elementwise_add_out = pattern->NewNode(elementwise_add_out_repr())
->assert_is_op_output("elementwise_add")
->AsOutput();
conv_op->LinksFrom({conv_in, conv_filter});
conv_out->LinksFrom({conv_op});
elementwise_add_op->LinksFrom({conv_out, elementwise_add_in_y})
.LinksTo({elementwise_add_out});
return elementwise_add_out;
}
} // namespace ir } // namespace ir
} // namespace framework } // namespace framework
} // namespace paddle } // namespace paddle
...@@ -716,6 +716,24 @@ struct ConvElementwiseadd2Act : public PatternBase { ...@@ -716,6 +716,24 @@ struct ConvElementwiseadd2Act : public PatternBase {
PATTERN_DECL_NODE(act_out); PATTERN_DECL_NODE(act_out);
}; };
// Conv + ElementwiseAdd
// This pattern should be used after ConvElementwiseadd2Act or
// ConvElementwiseadd pass
struct ConvElementwiseadd : public PatternBase {
ConvElementwiseadd(PDPattern* pattern, const std::string& name_scope)
: PatternBase(pattern, name_scope, "conv_elementwiseadd") {}
PDNode* operator()(PDNode* conv_in);
PATTERN_DECL_NODE(conv_op);
PATTERN_DECL_NODE(conv_out);
PATTERN_DECL_NODE(conv_filter);
PATTERN_DECL_NODE(elementwise_add_op);
PATTERN_DECL_NODE(elementwise_add_in_y);
PATTERN_DECL_NODE(elementwise_add_out);
};
} // namespace patterns } // namespace patterns
// Link two ir::Nodes from each other. // Link two ir::Nodes from each other.
......
...@@ -30,6 +30,14 @@ std::unique_ptr<Node> CreateNodeForTest(const std::string &name, ...@@ -30,6 +30,14 @@ std::unique_ptr<Node> CreateNodeForTest(const std::string &name,
return std::unique_ptr<Node>(new Node(name, type)); return std::unique_ptr<Node>(new Node(name, type));
} }
std::unique_ptr<Node> CreateNodeForTest(VarDesc *var_desc) {
return std::unique_ptr<Node>(new Node(var_desc));
}
std::unique_ptr<Node> CreateNodeForTest(OpDesc *op_desc) {
return std::unique_ptr<Node>(new Node(op_desc));
}
} // namespace ir } // namespace ir
} // namespace framework } // namespace framework
} // namespace paddle } // namespace paddle
...@@ -18,7 +18,6 @@ limitations under the License. */ ...@@ -18,7 +18,6 @@ limitations under the License. */
#include <typeindex> #include <typeindex>
#include <typeinfo> #include <typeinfo>
#include <vector> #include <vector>
#include "paddle/fluid/framework/op_desc.h" #include "paddle/fluid/framework/op_desc.h"
#include "paddle/fluid/framework/var_desc.h" #include "paddle/fluid/framework/var_desc.h"
#include "paddle/fluid/platform/macros.h" #include "paddle/fluid/platform/macros.h"
...@@ -125,6 +124,8 @@ class Node { ...@@ -125,6 +124,8 @@ class Node {
friend class Graph; friend class Graph;
friend std::unique_ptr<Node> CreateNodeForTest(const std::string& name, friend std::unique_ptr<Node> CreateNodeForTest(const std::string& name,
Node::Type type); Node::Type type);
friend std::unique_ptr<Node> CreateNodeForTest(VarDesc* var_desc);
friend std::unique_ptr<Node> CreateNodeForTest(OpDesc* op_desc);
explicit Node(const std::string& name, Type type) explicit Node(const std::string& name, Type type)
: name_(name), var_desc_(nullptr), op_desc_(nullptr), type_(type) {} : name_(name), var_desc_(nullptr), op_desc_(nullptr), type_(type) {}
...@@ -152,7 +153,9 @@ class Node { ...@@ -152,7 +153,9 @@ class Node {
std::unique_ptr<Node> CreateNodeForTest(const std::string& name, std::unique_ptr<Node> CreateNodeForTest(const std::string& name,
Node::Type type); Node::Type type);
std::unique_ptr<Node> CreateNodeForTest(VarDesc* var_desc);
std::unique_ptr<Node> CreateNodeForTest(OpDesc* op_desc);
} // namespace ir } // namespace ir
} // namespace framework } // namespace framework
} // namespace paddle } // namespace paddle
...@@ -278,7 +278,8 @@ std::shared_ptr<ngraph::runtime::Backend> NgraphEngine::backend_ = ...@@ -278,7 +278,8 @@ std::shared_ptr<ngraph::runtime::Backend> NgraphEngine::backend_ =
ngraph::runtime::Backend::create("CPU"); ngraph::runtime::Backend::create("CPU");
void NgraphEngine::GetNgInputShape(std::shared_ptr<OperatorBase> op) { void NgraphEngine::GetNgInputShape(std::shared_ptr<OperatorBase> op) {
op->RuntimeInferShape(scope_, place_); RuntimeContext ctx(op->Inputs(), op->Outputs(), scope_);
op->RuntimeInferShape(scope_, place_, ctx);
for (auto& var_name_item : op->Inputs()) { for (auto& var_name_item : op->Inputs()) {
for (auto& var_name : var_name_item.second) { for (auto& var_name : var_name_item.second) {
auto* var = scope_.FindVar(var_name); auto* var = scope_.FindVar(var_name);
......
...@@ -137,6 +137,23 @@ static LoD GetLoD(const Scope& scope, const std::string& name) { ...@@ -137,6 +137,23 @@ static LoD GetLoD(const Scope& scope, const std::string& name) {
} }
} }
RuntimeContext::RuntimeContext(const VariableNameMap& innames,
const VariableNameMap& outnames,
const Scope& scope) {
for (auto& var_name_item : innames) {
std::vector<Variable*>& input_vars = inputs[var_name_item.first];
for (auto& var_name : var_name_item.second) {
input_vars.push_back(scope.FindVar(var_name));
}
}
for (auto& var_name_item : outnames) {
std::vector<Variable*>& output_vars = outputs[var_name_item.first];
for (auto& var_name : var_name_item.second) {
output_vars.push_back(scope.FindVar(var_name));
}
}
}
void OperatorBase::Run(const Scope& scope, const platform::Place& place) { void OperatorBase::Run(const Scope& scope, const platform::Place& place) {
VLOG(4) << place << " " << DebugStringEx(&scope); VLOG(4) << place << " " << DebugStringEx(&scope);
if (platform::is_gpu_place(place)) { if (platform::is_gpu_place(place)) {
...@@ -412,11 +429,48 @@ bool ExecutionContext::HasOutput(const std::string& name) const { ...@@ -412,11 +429,48 @@ bool ExecutionContext::HasOutput(const std::string& name) const {
return var != nullptr; return var != nullptr;
} }
const Variable* ExecutionContext::InputVar(const std::string& name) const {
auto it = ctx_.inputs.find(name);
if (it == ctx_.inputs.end()) return nullptr;
PADDLE_ENFORCE_LE(it->second.size(), 1UL,
"Operator %s's input %s should contain only one variable.",
op_.Type(), name);
return it->second.empty() ? nullptr : it->second[0];
}
const Variable* ExecutionContext::LegacyInputVar(
const std::string& name) const {
auto ipt = op_.Input(name);
return ipt == kEmptyVarName ? nullptr : scope_.FindVar(ipt);
}
Variable* ExecutionContext::OutputVar(const std::string& name) const {
auto it = ctx_.outputs.find(name);
if (it == ctx_.outputs.end()) return nullptr;
PADDLE_ENFORCE_LE(it->second.size(), 1UL,
"Operator %s's output %s should contain only one variable.",
op_.Type(), name);
return it->second.empty() ? nullptr : it->second[0];
}
Variable* ExecutionContext::LegacyOutputVar(const std::string& name) const {
auto opt = op_.Output(name);
return opt == kEmptyVarName ? nullptr : scope_.FindVar(opt);
}
template <> template <>
const Tensor* ExecutionContext::Input<Tensor>(const std::string& name) const { const Tensor* ExecutionContext::Input<Tensor>(const std::string& name) const {
return Input<LoDTensor>(name); return Input<LoDTensor>(name);
} }
template <>
const Tensor* ExecutionContext::LegacyInput<Tensor>(
const std::string& name) const {
return LegacyInput<LoDTensor>(name);
}
template <> template <>
const std::vector<const Tensor*> ExecutionContext::MultiInput<Tensor>( const std::vector<const Tensor*> ExecutionContext::MultiInput<Tensor>(
const std::string& name) const { const std::string& name) const {
...@@ -441,6 +495,11 @@ Tensor* ExecutionContext::Output<Tensor>(const std::string& name) const { ...@@ -441,6 +495,11 @@ Tensor* ExecutionContext::Output<Tensor>(const std::string& name) const {
return Output<LoDTensor>(name); return Output<LoDTensor>(name);
} }
template <>
Tensor* ExecutionContext::LegacyOutput<Tensor>(const std::string& name) const {
return LegacyOutput<LoDTensor>(name);
}
template <> template <>
std::vector<Tensor*> ExecutionContext::MultiOutput<Tensor>( std::vector<Tensor*> ExecutionContext::MultiOutput<Tensor>(
const std::string& name) const { const std::string& name) const {
...@@ -477,23 +536,22 @@ bool OpSupportGPU(const std::string& op_type) { ...@@ -477,23 +536,22 @@ bool OpSupportGPU(const std::string& op_type) {
class RuntimeInferShapeContext : public InferShapeContext { class RuntimeInferShapeContext : public InferShapeContext {
public: public:
RuntimeInferShapeContext(const OperatorBase& op, const Scope& scope) RuntimeInferShapeContext(const OperatorBase& op, const Scope& scope,
: op_(op), scope_(scope) {} const RuntimeContext& ctx)
: op_(op), scope_(scope), ctx_(ctx) {}
bool HasInput(const std::string& name) const override { bool HasInput(const std::string& name) const override {
// has only one input // has only one input
const auto& ins = op_.Inputs(); const auto& ins = ctx_.inputs;
auto it = ins.find(name); auto it = ins.find(name);
if (it == ins.end()) { if (it == ins.end()) {
return false; return false;
} }
const auto& in = it->second; const auto& in = it->second;
if (in.size() == 0 || in[0] == kEmptyVarName) { if (in.size() == 0) return false;
return false;
}
PADDLE_ENFORCE_EQ(in.size(), 1UL, PADDLE_ENFORCE_EQ(in.size(), 1UL,
"Input %s should not have more than one inputs", name); "Input %s should not have more than one inputs", name);
return scope_.FindVar(in[0]) != nullptr; return in[0] != nullptr;
} }
bool HasOutput(const std::string& name) const override { bool HasOutput(const std::string& name) const override {
...@@ -678,6 +736,7 @@ class RuntimeInferShapeContext : public InferShapeContext { ...@@ -678,6 +736,7 @@ class RuntimeInferShapeContext : public InferShapeContext {
private: private:
const OperatorBase& op_; const OperatorBase& op_;
const Scope& scope_; const Scope& scope_;
const RuntimeContext& ctx_;
}; };
static void CheckTensorNANOrInf(const std::string& name, static void CheckTensorNANOrInf(const std::string& name,
...@@ -696,15 +755,15 @@ static void CheckTensorNANOrInf(const std::string& name, ...@@ -696,15 +755,15 @@ static void CheckTensorNANOrInf(const std::string& name,
} }
void OperatorWithKernel::RuntimeInferShape(const Scope& scope, void OperatorWithKernel::RuntimeInferShape(const Scope& scope,
const platform::Place& place) const { const platform::Place& place,
RuntimeInferShapeContext infer_shape_ctx(*this, scope); const RuntimeContext& ctx) const {
RuntimeInferShapeContext infer_shape_ctx(*this, scope, ctx);
this->InferShape(&infer_shape_ctx); this->InferShape(&infer_shape_ctx);
} }
void OperatorWithKernel::RunImpl(const Scope& scope, void OperatorWithKernel::RunImpl(const Scope& scope,
const platform::Place& place) const { const platform::Place& place) const {
RuntimeInferShapeContext infer_shape_ctx(*this, scope); RuntimeContext ctx(Inputs(), Outputs(), scope);
this->InferShape(&infer_shape_ctx);
platform::DeviceContextPool& pool = platform::DeviceContextPool::Instance(); platform::DeviceContextPool& pool = platform::DeviceContextPool::Instance();
auto* dev_ctx = pool.Get(place); auto* dev_ctx = pool.Get(place);
...@@ -718,15 +777,8 @@ void OperatorWithKernel::RunImpl(const Scope& scope, ...@@ -718,15 +777,8 @@ void OperatorWithKernel::RunImpl(const Scope& scope,
OpKernelMap& kernels = kernels_iter->second; OpKernelMap& kernels = kernels_iter->second;
// TODO(dzhwinter) : kernel fallback mechanism will be added when all the auto expected_kernel_key = this->GetExpectedKernelType(
// transform functions are ready. ExecutionContext(*this, scope, *dev_ctx, ctx));
// for (auto& candidate : kKernelPriority) {
// Do selection
// }
auto expected_kernel_key =
this->GetExpectedKernelType(ExecutionContext(*this, scope, *dev_ctx));
VLOG(3) << "expected_kernel_key:" << expected_kernel_key; VLOG(3) << "expected_kernel_key:" << expected_kernel_key;
auto kernel_iter = kernels.find(expected_kernel_key); auto kernel_iter = kernels.find(expected_kernel_key);
...@@ -748,7 +800,7 @@ void OperatorWithKernel::RunImpl(const Scope& scope, ...@@ -748,7 +800,7 @@ void OperatorWithKernel::RunImpl(const Scope& scope,
// do data transformScope &transfer_scope; // do data transformScope &transfer_scope;
std::vector<std::string> transfered_inplace_vars; std::vector<std::string> transfered_inplace_vars;
auto* transfer_scope = auto* transfer_scope =
TryTransferData(scope, expected_kernel_key, &transfered_inplace_vars); PrepareData(scope, expected_kernel_key, &transfered_inplace_vars, &ctx);
// exec scope is the scope that kernel actually executed on. // exec scope is the scope that kernel actually executed on.
const Scope& exec_scope = const Scope& exec_scope =
...@@ -758,7 +810,11 @@ void OperatorWithKernel::RunImpl(const Scope& scope, ...@@ -758,7 +810,11 @@ void OperatorWithKernel::RunImpl(const Scope& scope,
dev_ctx = pool.Get(expected_kernel_key.place_); dev_ctx = pool.Get(expected_kernel_key.place_);
} }
kernel_iter->second(ExecutionContext(*this, exec_scope, *dev_ctx)); RuntimeInferShapeContext infer_shape_ctx(*this, exec_scope, ctx);
this->InferShape(&infer_shape_ctx);
// TODO(panyx0718): ExecutionContext should only depend on RuntimeContext
// not Scope. Imperative mode only pass inputs and get outputs.
kernel_iter->second(ExecutionContext(*this, exec_scope, *dev_ctx, ctx));
if (!transfered_inplace_vars.empty()) { if (!transfered_inplace_vars.empty()) {
// there is inplace variable has been transfered. // there is inplace variable has been transfered.
...@@ -782,6 +838,7 @@ void OperatorWithKernel::RunImpl(const Scope& scope, ...@@ -782,6 +838,7 @@ void OperatorWithKernel::RunImpl(const Scope& scope,
} }
} }
} }
void OperatorWithKernel::TransferInplaceVarsBack( void OperatorWithKernel::TransferInplaceVarsBack(
const Scope& scope, const std::vector<std::string>& inplace_vars, const Scope& scope, const std::vector<std::string>& inplace_vars,
const Scope& transfer_scope) const { const Scope& transfer_scope) const {
...@@ -797,13 +854,19 @@ void OperatorWithKernel::TransferInplaceVarsBack( ...@@ -797,13 +854,19 @@ void OperatorWithKernel::TransferInplaceVarsBack(
} }
} }
Scope* OperatorWithKernel::TryTransferData( Scope* OperatorWithKernel::PrepareData(
const Scope& scope, const OpKernelType& expected_kernel_key, const Scope& scope, const OpKernelType& expected_kernel_key,
std::vector<std::string>* transfered_inplace_vars) const { std::vector<std::string>* transfered_inplace_vars,
RuntimeContext* ctx) const {
Scope* new_scope = nullptr; Scope* new_scope = nullptr;
for (auto& var_name_item : Inputs()) { for (auto& var_name_item : Inputs()) {
for (auto& var_name : var_name_item.second) { std::vector<Variable*>& input_vars = ctx->inputs[var_name_item.first];
for (size_t i = 0; i < var_name_item.second.size(); ++i) {
auto& var_name = var_name_item.second[i];
auto* var = scope.FindVar(var_name); auto* var = scope.FindVar(var_name);
input_vars[i] = var;
// Only tensor can be tranfer to another device. // Only tensor can be tranfer to another device.
if (var == nullptr || !VarIsTensor(*var)) { if (var == nullptr || !VarIsTensor(*var)) {
continue; continue;
...@@ -851,6 +914,7 @@ Scope* OperatorWithKernel::TryTransferData( ...@@ -851,6 +914,7 @@ Scope* OperatorWithKernel::TryTransferData(
} }
auto* trans_var = new_scope->Var(var_name); auto* trans_var = new_scope->Var(var_name);
input_vars[i] = trans_var;
Tensor out; Tensor out;
TransformData(expected_kernel_key, kernel_type_for_var, *tensor_in, &out); TransformData(expected_kernel_key, kernel_type_for_var, *tensor_in, &out);
......
...@@ -70,6 +70,15 @@ Tensor* GetMutableLoDTensorOrSelectedRowsValueFromVar(Variable* var); ...@@ -70,6 +70,15 @@ Tensor* GetMutableLoDTensorOrSelectedRowsValueFromVar(Variable* var);
class OperatorBase; class OperatorBase;
class ExecutionContext; class ExecutionContext;
class RuntimeContext {
public:
RuntimeContext(const VariableNameMap& innames,
const VariableNameMap& outnames, const Scope& scope);
VariableValueMap inputs;
VariableValueMap outputs;
};
/** /**
* OperatorBase has the basic elements that Net will call to do computation. * OperatorBase has the basic elements that Net will call to do computation.
* Only CreateOperator from OpRegistry will new Operator directly. User * Only CreateOperator from OpRegistry will new Operator directly. User
...@@ -129,7 +138,8 @@ class OperatorBase { ...@@ -129,7 +138,8 @@ class OperatorBase {
void SetIsCalledByExecutor(bool x) { run_by_executor_ = x; } void SetIsCalledByExecutor(bool x) { run_by_executor_ = x; }
virtual void RuntimeInferShape(const Scope& scope, virtual void RuntimeInferShape(const Scope& scope,
const platform::Place& place) const {} const platform::Place& place,
const RuntimeContext& ctx) const {}
protected: protected:
std::string type_; std::string type_;
...@@ -156,8 +166,9 @@ class OperatorBase { ...@@ -156,8 +166,9 @@ class OperatorBase {
class ExecutionContext { class ExecutionContext {
public: public:
ExecutionContext(const OperatorBase& op, const Scope& scope, ExecutionContext(const OperatorBase& op, const Scope& scope,
const platform::DeviceContext& device_context) const platform::DeviceContext& device_context,
: op_(op), scope_(scope), device_context_(device_context) {} const RuntimeContext& ctx)
: op_(op), scope_(scope), device_context_(device_context), ctx_(ctx) {}
const OperatorBase& op() const { return op_; } const OperatorBase& op() const { return op_; }
...@@ -180,15 +191,9 @@ class ExecutionContext { ...@@ -180,15 +191,9 @@ class ExecutionContext {
return op_.Outputs(name).size(); return op_.Outputs(name).size();
} }
const Variable* InputVar(const std::string& name) const { const Variable* InputVar(const std::string& name) const;
auto ipt = op_.Input(name);
return ipt == kEmptyVarName ? nullptr : scope_.FindVar(ipt);
}
Variable* OutputVar(const std::string& name) const { Variable* OutputVar(const std::string& name) const;
auto opt = op_.Output(name);
return opt == kEmptyVarName ? nullptr : scope_.FindVar(opt);
}
const std::vector<const Variable*> MultiInputVar( const std::vector<const Variable*> MultiInputVar(
const std::string& name) const { const std::string& name) const {
...@@ -227,6 +232,22 @@ class ExecutionContext { ...@@ -227,6 +232,22 @@ class ExecutionContext {
return var == nullptr ? nullptr : var->GetMutable<T>(); return var == nullptr ? nullptr : var->GetMutable<T>();
} }
template <typename T>
const T* LegacyInput(const std::string& name) const {
auto* var = LegacyInputVar(name);
return var == nullptr ? nullptr : &var->Get<T>();
}
template <typename T>
T* LegacyOutput(const std::string& name) const {
auto var = LegacyOutputVar(name);
return var == nullptr ? nullptr : var->GetMutable<T>();
}
const Variable* LegacyInputVar(const std::string& name) const;
Variable* LegacyOutputVar(const std::string& name) const;
template <typename T> template <typename T>
const std::vector<const T*> MultiInput(const std::string& name) const { const std::vector<const T*> MultiInput(const std::string& name) const {
auto names = op_.Inputs(name); auto names = op_.Inputs(name);
...@@ -286,11 +307,16 @@ class ExecutionContext { ...@@ -286,11 +307,16 @@ class ExecutionContext {
const OperatorBase& op_; const OperatorBase& op_;
const Scope& scope_; const Scope& scope_;
const platform::DeviceContext& device_context_; const platform::DeviceContext& device_context_;
const RuntimeContext& ctx_;
}; };
template <> template <>
const Tensor* ExecutionContext::Input<Tensor>(const std::string& name) const; const Tensor* ExecutionContext::Input<Tensor>(const std::string& name) const;
template <>
const Tensor* ExecutionContext::LegacyInput<Tensor>(
const std::string& name) const;
template <> template <>
const std::vector<const Tensor*> ExecutionContext::MultiInput<Tensor>( const std::vector<const Tensor*> ExecutionContext::MultiInput<Tensor>(
const std::string& name) const; const std::string& name) const;
...@@ -298,6 +324,9 @@ const std::vector<const Tensor*> ExecutionContext::MultiInput<Tensor>( ...@@ -298,6 +324,9 @@ const std::vector<const Tensor*> ExecutionContext::MultiInput<Tensor>(
template <> template <>
Tensor* ExecutionContext::Output<Tensor>(const std::string& name) const; Tensor* ExecutionContext::Output<Tensor>(const std::string& name) const;
template <>
Tensor* ExecutionContext::LegacyOutput<Tensor>(const std::string& name) const;
template <> template <>
std::vector<Tensor*> ExecutionContext::MultiOutput<Tensor>( std::vector<Tensor*> ExecutionContext::MultiOutput<Tensor>(
const std::string& name) const; const std::string& name) const;
...@@ -350,8 +379,8 @@ class OperatorWithKernel : public OperatorBase { ...@@ -350,8 +379,8 @@ class OperatorWithKernel : public OperatorBase {
OpInfoMap::Instance().Get(Type()).infer_shape_(ctx); OpInfoMap::Instance().Get(Type()).infer_shape_(ctx);
} }
void RuntimeInferShape(const Scope& scope, void RuntimeInferShape(const Scope& scope, const platform::Place& place,
const platform::Place& place) const override; const RuntimeContext& ctx) const override;
protected: protected:
virtual OpKernelType GetExpectedKernelType(const ExecutionContext& ctx) const; virtual OpKernelType GetExpectedKernelType(const ExecutionContext& ctx) const;
...@@ -371,9 +400,10 @@ class OperatorWithKernel : public OperatorBase { ...@@ -371,9 +400,10 @@ class OperatorWithKernel : public OperatorBase {
* *
* * transfered_inplace_vars is a output vector. * * transfered_inplace_vars is a output vector.
*/ */
Scope* TryTransferData( Scope* PrepareData(const Scope& scope,
const Scope& scope, const OpKernelType& expected_kernel_key, const OpKernelType& expected_kernel_key,
std::vector<std::string>* transfered_inplace_vars) const; std::vector<std::string>* transfered_inplace_vars,
RuntimeContext* ctx) const;
void TransferInplaceVarsBack(const Scope& scope, void TransferInplaceVarsBack(const Scope& scope,
const std::vector<std::string>& inplace_vars, const std::vector<std::string>& inplace_vars,
......
...@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and ...@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and
limitations under the License. */ limitations under the License. */
#include "paddle/fluid/framework/parallel_executor.h" #include "paddle/fluid/framework/parallel_executor.h"
#include <algorithm>
#include <string> #include <string>
#include <tuple> #include <tuple>
#include <vector> #include <vector>
...@@ -93,6 +94,7 @@ class ParallelExecutorPrivate { ...@@ -93,6 +94,7 @@ class ParallelExecutorPrivate {
} }
} }
BuildStrategy build_strategy_;
std::vector<platform::Place> places_; std::vector<platform::Place> places_;
std::vector<Scope *> local_scopes_; std::vector<Scope *> local_scopes_;
Scope *global_scope_; // not owned Scope *global_scope_; // not owned
...@@ -169,6 +171,14 @@ std::unique_ptr<ir::Graph> ParallelExecutorPrivate::PrepareGCAndRefCnts( ...@@ -169,6 +171,14 @@ std::unique_ptr<ir::Graph> ParallelExecutorPrivate::PrepareGCAndRefCnts(
eager_deletion_pass->SetNotOwned(details::kAllPlaces, &places_); eager_deletion_pass->SetNotOwned(details::kAllPlaces, &places_);
graph = eager_deletion_pass->Apply(std::move(graph)); graph = eager_deletion_pass->Apply(std::move(graph));
VLOG(10) << "EagerDeletionPass Applied"; VLOG(10) << "EagerDeletionPass Applied";
if (build_strategy_.memory_early_delete_) {
auto early_delete_pass =
ir::PassRegistry::Instance().Get("memory_early_delete_pass");
early_delete_pass->SetNotOwned(details::kGarbageCollector, &gcs_);
graph = early_delete_pass->Apply(std::move(graph));
}
VLOG(10) << "MemoryEarlyDeletePass Applied.";
} }
return graph; return graph;
...@@ -189,6 +199,7 @@ ParallelExecutor::ParallelExecutor( ...@@ -189,6 +199,7 @@ ParallelExecutor::ParallelExecutor(
: member_(new ParallelExecutorPrivate(places)) { : member_(new ParallelExecutorPrivate(places)) {
member_->global_scope_ = scope; member_->global_scope_ = scope;
member_->use_cuda_ = exec_strategy.use_cuda_; member_->use_cuda_ = exec_strategy.use_cuda_;
member_->build_strategy_ = build_strategy;
member_->use_all_reduce_ = member_->use_all_reduce_ =
build_strategy.reduce_ == BuildStrategy::ReduceStrategy::kAllReduce; build_strategy.reduce_ == BuildStrategy::ReduceStrategy::kAllReduce;
...@@ -245,7 +256,6 @@ ParallelExecutor::ParallelExecutor( ...@@ -245,7 +256,6 @@ ParallelExecutor::ParallelExecutor(
build_strategy.Apply(main_program, member_->places_, loss_var_name, build_strategy.Apply(main_program, member_->places_, loss_var_name,
params, member_->local_scopes_, member_->use_cuda_); params, member_->local_scopes_, member_->use_cuda_);
#endif #endif
auto max_memory_size = GetEagerDeletionThreshold(); auto max_memory_size = GetEagerDeletionThreshold();
if (max_memory_size >= 0) { if (max_memory_size >= 0) {
graph = member_->PrepareGCAndRefCnts(std::move(graph), graph = member_->PrepareGCAndRefCnts(std::move(graph),
...@@ -280,10 +290,12 @@ ParallelExecutor::ParallelExecutor( ...@@ -280,10 +290,12 @@ ParallelExecutor::ParallelExecutor(
if (exec_strategy.type_ == ExecutionStrategy::kDefault) { if (exec_strategy.type_ == ExecutionStrategy::kDefault) {
member_->executor_.reset(new details::ThreadedSSAGraphExecutor( member_->executor_.reset(new details::ThreadedSSAGraphExecutor(
exec_strategy, member_->local_scopes_, places, std::move(graph))); exec_strategy, member_->local_scopes_, member_->places_,
std::move(graph)));
} else { } else {
member_->executor_.reset(new details::FastThreadedSSAGraphExecutor( member_->executor_.reset(new details::FastThreadedSSAGraphExecutor(
exec_strategy, member_->local_scopes_, places, std::move(graph))); exec_strategy, member_->local_scopes_, member_->places_,
std::move(graph)));
} }
member_->executor_.reset(new details::ScopeBufferedSSAGraphExecutor( member_->executor_.reset(new details::ScopeBufferedSSAGraphExecutor(
...@@ -423,5 +435,6 @@ ParallelExecutor::~ParallelExecutor() { ...@@ -423,5 +435,6 @@ ParallelExecutor::~ParallelExecutor() {
} // namespace framework } // namespace framework
} // namespace paddle } // namespace paddle
USE_PASS(memory_early_delete_pass);
USE_PASS(reference_count_pass); USE_PASS(reference_count_pass);
USE_PASS(eager_deletion_pass); USE_PASS(eager_deletion_pass);
...@@ -74,6 +74,22 @@ TEST(Tensor, MutableData) { ...@@ -74,6 +74,22 @@ TEST(Tensor, MutableData) {
p2 = src_tensor.mutable_data<float>(framework::make_ddim({2, 2}), p2 = src_tensor.mutable_data<float>(framework::make_ddim({2, 2}),
platform::CPUPlace()); platform::CPUPlace());
EXPECT_EQ(p1, p2); EXPECT_EQ(p1, p2);
float* p3 = nullptr;
float* p4 = nullptr;
// set src_tensor a different type but smaller size.
// memory block is supposed to be unchanged.
auto* tmp = src_tensor.mutable_data<uint8_t>(framework::make_ddim({2, 2}),
platform::CPUPlace());
p3 = reinterpret_cast<float*>(tmp);
EXPECT_EQ(p1, p3);
// set src_tensor a different type but bigger size.
// memory block is supposed to be changed.
auto* tmp2 = src_tensor.mutable_data<double>(
framework::make_ddim({2, 2, 3}), platform::CPUPlace());
p4 = reinterpret_cast<float*>(tmp2);
EXPECT_NE(p1, p4);
} }
// Not sure if it's desired, but currently, Tensor type can be changed. // Not sure if it's desired, but currently, Tensor type can be changed.
{ {
......
...@@ -28,8 +28,11 @@ class OperatorBase; ...@@ -28,8 +28,11 @@ class OperatorBase;
class OpDesc; class OpDesc;
class InferShapeContext; class InferShapeContext;
class BlockDesc; class BlockDesc;
class Variable;
using VariableNameMap = std::map<std::string, std::vector<std::string>>; using VariableNameMap = std::map<std::string, std::vector<std::string>>;
// TODO(panyx0718): Replace vector with something like gtl::Vector.
using VariableValueMap = std::map<std::string, std::vector<Variable*>>;
// The order should be as same as framework.proto // The order should be as same as framework.proto
using Attribute = using Attribute =
......
...@@ -63,7 +63,6 @@ std::unique_ptr<framework::ir::Graph> analysis::TensorRtSubgraphPass::ApplyImpl( ...@@ -63,7 +63,6 @@ std::unique_ptr<framework::ir::Graph> analysis::TensorRtSubgraphPass::ApplyImpl(
void TensorRtSubgraphPass::CreateTensorRTOp(framework::ir::Node *node, void TensorRtSubgraphPass::CreateTensorRTOp(framework::ir::Node *node,
Graph *graph) const { Graph *graph) const {
auto *op_desc = node->Op(); auto *op_desc = node->Op();
static int counter{0};
auto &subgraph = *Agent(node).subgraph(); auto &subgraph = *Agent(node).subgraph();
PADDLE_ENFORCE(!subgraph.empty()); PADDLE_ENFORCE(!subgraph.empty());
...@@ -192,8 +191,6 @@ void TensorRtSubgraphPass::CreateTensorRTOp(framework::ir::Node *node, ...@@ -192,8 +191,6 @@ void TensorRtSubgraphPass::CreateTensorRTOp(framework::ir::Node *node,
block_desc.Proto()->SerializeAsString()); block_desc.Proto()->SerializeAsString());
SetAttr(op_desc->Proto(), "max_batch_size", Get<int>("max_batch_size")); SetAttr(op_desc->Proto(), "max_batch_size", Get<int>("max_batch_size"));
SetAttr(op_desc->Proto(), "workspace_size", Get<int>("workspace_size")); SetAttr(op_desc->Proto(), "workspace_size", Get<int>("workspace_size"));
SetAttr(op_desc->Proto(), "engine_uniq_key",
"trt-" + std::to_string(counter++));
SetAttr(op_desc->Proto(), "parameters", ExtractParameters(graph->Nodes())); SetAttr(op_desc->Proto(), "parameters", ExtractParameters(graph->Nodes()));
SetAttr(op_desc->Proto(), "output_name_mapping", output_mapping); SetAttr(op_desc->Proto(), "output_name_mapping", output_mapping);
} }
......
...@@ -122,6 +122,7 @@ class GpuPassStrategy : public PassStrategy { ...@@ -122,6 +122,7 @@ class GpuPassStrategy : public PassStrategy {
"conv_bn_fuse_pass", // "conv_bn_fuse_pass", //
"conv_elementwise_add_act_fuse_pass", // "conv_elementwise_add_act_fuse_pass", //
"conv_elementwise_add2_act_fuse_pass", // "conv_elementwise_add2_act_fuse_pass", //
"conv_elementwise_add_fuse_pass", //
}); });
} }
......
...@@ -103,6 +103,7 @@ class OpConverter { ...@@ -103,6 +103,7 @@ class OpConverter {
void ConvertBlock(const framework::proto::BlockDesc& block, void ConvertBlock(const framework::proto::BlockDesc& block,
const std::unordered_set<std::string>& parameters, const std::unordered_set<std::string>& parameters,
const framework::Scope& scope, TensorRTEngine* engine) { const framework::Scope& scope, TensorRTEngine* engine) {
std::unique_lock<std::mutex> lk(mut_);
for (int i = 0; i < block.ops_size(); i++) { for (int i = 0; i < block.ops_size(); i++) {
const auto& op = block.ops(i); const auto& op = block.ops(i);
ConvertOp(op, parameters, scope, engine); ConvertOp(op, parameters, scope, engine);
...@@ -125,6 +126,7 @@ class OpConverter { ...@@ -125,6 +126,7 @@ class OpConverter {
std::unordered_map<std::string, OpConverter*> converters_; std::unordered_map<std::string, OpConverter*> converters_;
// fluid inference scope // fluid inference scope
framework::Scope* scope_{nullptr}; framework::Scope* scope_{nullptr};
std::mutex mut_;
}; };
} // namespace tensorrt } // namespace tensorrt
......
...@@ -30,6 +30,13 @@ function(inference_analysis_api_test_with_fake_data target install_dir filename ...@@ -30,6 +30,13 @@ function(inference_analysis_api_test_with_fake_data target install_dir filename
ARGS --infer_model=${install_dir}/model) ARGS --infer_model=${install_dir}/model)
endfunction() endfunction()
function(inference_analysis_api_test_with_refer_result target install_dir filename)
inference_analysis_test(${target} SRCS ${filename}
EXTRA_DEPS ${INFERENCE_EXTRA_DEPS}
ARGS --infer_model=${install_dir}/model --infer_data=${install_dir}/data.txt
--refer_result=${install_dir}/result.txt)
endfunction()
# RNN1 # RNN1
if(NOT APPLE AND WITH_MKLML) if(NOT APPLE AND WITH_MKLML)
set(RNN1_INSTALL_DIR "${INFERENCE_DEMO_INSTALL_DIR}/rnn1") set(RNN1_INSTALL_DIR "${INFERENCE_DEMO_INSTALL_DIR}/rnn1")
...@@ -83,14 +90,21 @@ set(OCR_INSTALL_DIR "${INFERENCE_DEMO_INSTALL_DIR}/ocr") ...@@ -83,14 +90,21 @@ set(OCR_INSTALL_DIR "${INFERENCE_DEMO_INSTALL_DIR}/ocr")
if (NOT EXISTS ${OCR_INSTALL_DIR}) if (NOT EXISTS ${OCR_INSTALL_DIR})
inference_download_and_uncompress(${OCR_INSTALL_DIR} "http://paddlemodels.cdn.bcebos.com/" "inference-vis-demos%2Focr.tar.gz") inference_download_and_uncompress(${OCR_INSTALL_DIR} "http://paddlemodels.cdn.bcebos.com/" "inference-vis-demos%2Focr.tar.gz")
endif() endif()
inference_analysis_api_test(test_analyzer_ocr ${OCR_INSTALL_DIR} analyzer_vis_tester.cc) inference_analysis_api_test_with_refer_result(test_analyzer_ocr ${OCR_INSTALL_DIR} analyzer_vis_tester.cc)
# mobilenet with transpose op
set(MOBILENET_INSTALL_DIR "${INFERENCE_DEMO_INSTALL_DIR}/mobilenet")
if (NOT EXISTS ${MOBILENET_INSTALL_DIR})
inference_download_and_uncompress(${MOBILENET_INSTALL_DIR} "http://paddlemodels.cdn.bcebos.com/" "inference-vis-demos%2Fmobilenet.tar.gz")
endif()
inference_analysis_api_test_with_refer_result(test_analyzer_mobilenet_transpose ${MOBILENET_INSTALL_DIR} analyzer_vis_tester.cc)
# resnet50 # resnet50
inference_analysis_api_test_with_fake_data(test_analyzer_resnet50 inference_analysis_api_test_with_fake_data(test_analyzer_resnet50
"${INFERENCE_DEMO_INSTALL_DIR}/resnet50" analyzer_resnet50_tester.cc "resnet50_model.tar.gz") "${INFERENCE_DEMO_INSTALL_DIR}/resnet50" analyzer_resnet50_tester.cc "resnet50_model.tar.gz")
# mobilenet with depthwise_conv op # mobilenet with depthwise_conv op
inference_analysis_api_test_with_fake_data(test_analyzer_mobilenet inference_analysis_api_test_with_fake_data(test_analyzer_mobilenet_depthwise_conv
"${INFERENCE_DEMO_INSTALL_DIR}/mobilenet_depthwise_conv" analyzer_resnet50_tester.cc "mobilenet_model.tar.gz") "${INFERENCE_DEMO_INSTALL_DIR}/mobilenet_depthwise_conv" analyzer_resnet50_tester.cc "mobilenet_model.tar.gz")
# anakin # anakin
......
...@@ -93,18 +93,20 @@ void profile(bool use_mkldnn = false) { ...@@ -93,18 +93,20 @@ void profile(bool use_mkldnn = false) {
SetInput(&input_slots_all); SetInput(&input_slots_all);
TestPrediction(reinterpret_cast<const PaddlePredictor::Config *>(&cfg), TestPrediction(reinterpret_cast<const PaddlePredictor::Config *>(&cfg),
input_slots_all, &outputs, FLAGS_num_threads); input_slots_all, &outputs, FLAGS_num_threads);
if (FLAGS_num_threads == 1 && !FLAGS_test_all_data) { if (FLAGS_num_threads == 1 && !FLAGS_test_all_data) {
const float ocr_result_data[] = { std::string line;
5.273636460856323538e-08, 3.296741795111302054e-07, std::ifstream file(FLAGS_refer_result);
1.873261190610264748e-08, 3.403730275408634043e-08, std::getline(file, line);
3.383312474625199684e-08}; auto refer = ProcessALine(line);
PADDLE_ENFORCE_EQ(outputs.size(), 1UL); file.close();
size_t size = GetSize(outputs[0]);
PADDLE_ENFORCE_GT(size, 0); auto &output = outputs.front();
float *result = static_cast<float *>(outputs[0].data.data()); size_t numel = output.data.length() / PaddleDtypeSize(output.dtype);
for (size_t i = 0; i < std::min(5UL, size); i++) { CHECK_EQ(numel, refer.data.size());
EXPECT_NEAR(result[i], ocr_result_data[i], 1e-3); for (size_t i = 0; i < numel; ++i) {
CHECK_LT(
fabs(static_cast<float *>(output.data.data())[i] - refer.data[i]),
1e-5);
} }
} }
} }
......
...@@ -36,6 +36,7 @@ ...@@ -36,6 +36,7 @@
DEFINE_string(model_name, "", "model name"); DEFINE_string(model_name, "", "model name");
DEFINE_string(infer_model, "", "model path"); DEFINE_string(infer_model, "", "model path");
DEFINE_string(infer_data, "", "data file"); DEFINE_string(infer_data, "", "data file");
DEFINE_string(refer_result, "", "reference result for comparison");
DEFINE_int32(batch_size, 1, "batch size."); DEFINE_int32(batch_size, 1, "batch size.");
DEFINE_int32(repeat, 1, "Running the inference program repeat times."); DEFINE_int32(repeat, 1, "Running the inference program repeat times.");
DEFINE_bool(test_all_data, false, "Test the all dataset in data file."); DEFINE_bool(test_all_data, false, "Test the all dataset in data file.");
......
...@@ -64,9 +64,7 @@ endif() ...@@ -64,9 +64,7 @@ endif()
set(COMMON_OP_DEPS ${OP_HEADER_DEPS}) set(COMMON_OP_DEPS ${OP_HEADER_DEPS})
set(COMMON_OP_DEPS ${COMMON_OP_DEPS} selected_rows_functor selected_rows lod_tensor maxouting unpooling pooling lod_rank_table context_project sequence_pooling executor) set(COMMON_OP_DEPS ${COMMON_OP_DEPS} selected_rows_functor selected_rows lod_tensor maxouting unpooling pooling lod_rank_table context_project sequence_pooling executor)
if (NOT WIN32) set(COMMON_OP_DEPS ${COMMON_OP_DEPS} dynload_warpctc)
set(COMMON_OP_DEPS ${COMMON_OP_DEPS} dynload_warpctc)
endif()
set(COMMON_OP_DEPS ${COMMON_OP_DEPS} sequence_padding sequence_scale cos_sim_functor memory jit_kernel concat_and_split cross_entropy softmax vol2col im2col sampler) set(COMMON_OP_DEPS ${COMMON_OP_DEPS} sequence_padding sequence_scale cos_sim_functor memory jit_kernel concat_and_split cross_entropy softmax vol2col im2col sampler)
set(COMMON_OP_DEPS ${COMMON_OP_DEPS} sequence2batch lstm_compute matrix_bit_code gru_compute activation_functions) set(COMMON_OP_DEPS ${COMMON_OP_DEPS} sequence2batch lstm_compute matrix_bit_code gru_compute activation_functions)
if (WITH_GPU) if (WITH_GPU)
......
...@@ -122,7 +122,8 @@ class BeamSearchDecodeOp : public framework::OperatorBase { ...@@ -122,7 +122,8 @@ class BeamSearchDecodeOp : public framework::OperatorBase {
platform::DeviceContextPool& pool = platform::DeviceContextPool::Instance(); platform::DeviceContextPool& pool = platform::DeviceContextPool::Instance();
auto& dev_ctx = *pool.Get(dev_place); auto& dev_ctx = *pool.Get(dev_place);
framework::ExecutionContext ctx(*this, scope, dev_ctx); framework::RuntimeContext run_ctx(Inputs(), Outputs(), scope);
framework::ExecutionContext ctx(*this, scope, dev_ctx, run_ctx);
const LoDTensorArray* ids = ctx.Input<LoDTensorArray>("Ids"); const LoDTensorArray* ids = ctx.Input<LoDTensorArray>("Ids");
const LoDTensorArray* scores = ctx.Input<LoDTensorArray>("Scores"); const LoDTensorArray* scores = ctx.Input<LoDTensorArray>("Scores");
......
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License. */
#include <vector>
#include "paddle/fluid/framework/executor.h"
#include "paddle/fluid/framework/op_registry.h"
#include "paddle/fluid/framework/threadpool.h"
#include "paddle/fluid/operators/detail/safe_ref.h"
namespace paddle {
namespace operators {
static constexpr char kInputs[] = "inputs";
static constexpr char kParameters[] = "parameters";
static constexpr char kPlaces[] = "places";
static constexpr char kOutputs[] = "outputs";
static constexpr char kParallelScopes[] = "parallel_scopes";
static constexpr char kParallelBlock[] = "sub_block";
static constexpr char kUseNCCL[] = "use_nccl";
using LoDTensor = framework::LoDTensor;
using SelectedRows = framework::SelectedRows;
static void SplitTensorAndMoveTensorToScopes(
const framework::Scope &scope, std::vector<framework::Scope *> *sub_scopes,
const std::vector<platform::Place> &places,
const std::vector<std::string> &names) {
size_t num_sub_scopes = 0;
for (auto &argu : names) {
const auto &tensor =
detail::Ref(scope.FindVar(argu),
"Cannot find variable %s in the parent scope", argu)
.Get<LoDTensor>();
auto lod_tensors = tensor.SplitLoDTensor(places);
for (auto &lod : lod_tensors) {
VLOG(3) << lod.dims();
}
if (num_sub_scopes == 0) {
num_sub_scopes = lod_tensors.size();
} else {
PADDLE_ENFORCE_EQ(num_sub_scopes, lod_tensors.size());
}
PADDLE_ENFORCE_NE(num_sub_scopes, 0);
if (sub_scopes->size() == 0) {
sub_scopes->reserve(num_sub_scopes);
for (size_t i = 0; i < num_sub_scopes; ++i) {
sub_scopes->emplace_back(&scope.NewScope());
}
}
for (size_t i = 0; i < lod_tensors.size(); ++i) {
*detail::Ref(sub_scopes->at(i)->Var(argu),
"Cannot find variable in the sub-scope", argu)
.GetMutable<LoDTensor>() = lod_tensors[i];
}
}
}
inline void CopyOrShare(const framework::Variable &src,
const platform::Place &dst_place,
framework::Variable *dst) {
if (src.IsType<LoDTensor>()) {
if (src.Get<LoDTensor>().place() == dst_place) {
dst->GetMutable<LoDTensor>()->ShareDataWith(src.Get<LoDTensor>());
dst->GetMutable<LoDTensor>()->set_lod(src.Get<LoDTensor>().lod());
} else {
TensorCopy(src.Get<LoDTensor>(), dst_place, dst->GetMutable<LoDTensor>());
}
} else if (src.IsType<SelectedRows>()) {
auto &src_sr = src.Get<SelectedRows>();
auto *dst_sr = dst->GetMutable<SelectedRows>();
dst_sr->set_height(src_sr.height());
if (src_sr.value().place() == dst_place) {
dst_sr->mutable_value()->ShareDataWith(src_sr.value());
dst_sr->set_rows(src_sr.rows());
} else {
TensorCopy(src_sr.value(), dst_place, dst_sr->mutable_value());
}
} else {
PADDLE_THROW("Expect LoDTensor/SelectedRows, get %s", src.Type().name());
}
}
void WaitOnPlace(const platform::Place place) {
platform::DeviceContextPool &pool = platform::DeviceContextPool::Instance();
auto &dev_ctx = *pool.Get(place);
dev_ctx.Wait();
}
void WaitOnPlaces(const std::vector<platform::Place> places) {
platform::DeviceContextPool &pool = platform::DeviceContextPool::Instance();
for (auto &place : places) {
auto &dev_ctx = *pool.Get(place);
dev_ctx.Wait();
}
}
class ParallelDoOp : public framework::OperatorBase {
public:
ParallelDoOp(const std::string &type,
const framework::VariableNameMap &inputs,
const framework::VariableNameMap &outputs,
const framework::AttributeMap &attrs)
: framework::OperatorBase(type, inputs, outputs, attrs) {}
private:
void RunImpl(const framework::Scope &scope,
const platform::Place &place) const override {
// get device context from pool
platform::DeviceContextPool &pool = platform::DeviceContextPool::Instance();
auto &dev_ctx = *pool.Get(place);
auto *block = Attr<framework::BlockDesc *>(kParallelBlock);
auto *program = block->Program();
auto &places = scope.FindVar(Input(kPlaces))->Get<platform::PlaceList>();
auto &sub_scopes = *scope.FindVar(Output(kParallelScopes))
->GetMutable<std::vector<framework::Scope *>>();
// split input
SplitTensorAndMoveTensorToScopes(scope, &sub_scopes, places,
Inputs(kInputs));
// copy parameter
for (auto &param : Inputs(kParameters)) {
PADDLE_ENFORCE(scope.FindVar(param)->IsType<LoDTensor>(),
"Only support parameter type as LoDTensor");
auto &src = scope.FindVar(param)->Get<LoDTensor>();
auto *sub_scope0 = sub_scopes[0];
auto *dst0 = sub_scope0->Var(param)->GetMutable<LoDTensor>();
dst0->ShareDataWith(src);
for (size_t i = 1; i < sub_scopes.size(); ++i) {
auto &place = places[i];
auto *sub_scope = sub_scopes[i];
auto *dst = sub_scope->Var(param)->GetMutable<LoDTensor>();
framework::TensorCopy(src, place, dst);
}
}
WaitOnPlaces(places);
std::vector<std::future<void>> workers;
workers.reserve(places.size());
for (size_t place_idx = 0; place_idx < sub_scopes.size(); ++place_idx) {
auto &place = places[place_idx];
auto *cur_scope = sub_scopes[place_idx];
workers.emplace_back(framework::Async([program, cur_scope, place, block] {
framework::Executor executor(place);
executor.Run(*program, cur_scope, block->ID(),
false /*create_local_scope*/);
}));
}
for (auto &worker : workers) {
worker.wait();
}
WaitOnPlaces(places);
// merge output
for (auto &o_name : Outputs(kOutputs)) {
std::vector<const framework::LoDTensor *> lod_tensors;
lod_tensors.reserve(sub_scopes.size());
for (auto *sub_scope : sub_scopes) {
lod_tensors.emplace_back(&sub_scope->FindVar(o_name)->Get<LoDTensor>());
}
auto *lod_tensor_to_be_merged =
scope.FindVar(o_name)->GetMutable<LoDTensor>();
lod_tensor_to_be_merged->MergeLoDTensor(lod_tensors, dev_ctx.GetPlace());
}
WaitOnPlaces(places);
}
};
class ParallelDoOpProtoMaker : public framework::OpProtoAndCheckerMaker {
public:
void Make() override {
AddInput(kInputs, "").AsDuplicable();
AddInput(kParameters, "").AsDuplicable();
AddInput(kPlaces, "");
AddOutput(kOutputs, "").AsDuplicable();
AddOutput(kParallelScopes, "");
AddAttr<framework::BlockDesc *>(kParallelBlock, "");
AddAttr<bool>(kUseNCCL, "true if we use nccl on backward")
.SetDefault(false);
AddComment(R"DOC(
ParallelDo Operator.
)DOC");
}
};
class ParallelDoGradOp : public framework::OperatorBase {
public:
ParallelDoGradOp(const std::string &type,
const framework::VariableNameMap &inputs,
const framework::VariableNameMap &outputs,
const framework::AttributeMap &attrs)
: framework::OperatorBase(type, inputs, outputs, attrs) {}
private:
void RunImpl(const framework::Scope &scope,
const platform::Place &place) const override {
auto *block = Attr<framework::BlockDesc *>(kParallelBlock);
auto *program = block->Program();
auto &sub_scopes = scope.FindVar(Input(kParallelScopes))
->Get<std::vector<framework::Scope *>>();
auto &places = scope.FindVar(Input(kPlaces))->Get<platform::PlaceList>();
// feed output@grad
SplitTensorAndMoveTensorToScopes(
scope, const_cast<std::vector<framework::Scope *> *>(&sub_scopes),
places, Inputs(framework::GradVarName(kOutputs)));
WaitOnPlaces(places);
// exe run
std::vector<std::future<void>> workers;
for (size_t i = 0; i < sub_scopes.size(); ++i) {
auto &place = places[i];
auto *cur_scope = sub_scopes[i];
// execute
workers.emplace_back(framework::Async([program, cur_scope, place, block] {
framework::Executor executor(place);
executor.Run(*program, cur_scope, block->ID(),
false /*create_local_scope*/);
}));
}
for (auto &worker : workers) {
worker.wait();
}
WaitOnPlaces(places);
// NCCL allreduce op will be added by backward,
// so no need to explicitly accumulate grad
if (!(Attr<bool>(kUseNCCL))) {
AccumulateGrad(scope, place, sub_scopes, places);
} else {
for (auto &place : places) {
PADDLE_ENFORCE(platform::is_gpu_place(place),
"NCCL only supports cuda place");
}
}
for (auto &s : Outputs(framework::GradVarName(kParameters))) {
if (s == framework::kEmptyVarName) {
continue;
}
VLOG(3) << "Moving " << s;
CopyOrShare(*sub_scopes[0]->FindVar(s), place, scope.FindVar(s));
}
WaitOnPlaces(places);
}
void AccumulateGrad(const framework::Scope &scope,
const platform::Place &place,
const std::vector<framework::Scope *> &sub_scopes,
const platform::PlaceList &places) const {
for (auto &s : Outputs(framework::GradVarName(kParameters))) {
if (s == framework::kEmptyVarName) {
continue;
}
VLOG(3) << "Accumulating " << s;
if (s == framework::kEmptyVarName) continue;
std::string tmp_name;
auto *tmp = sub_scopes[0]->Var(&tmp_name);
for (size_t i = 1; i < sub_scopes.size(); ++i) {
CopyOrShare(*sub_scopes[i]->FindVar(s), places[0], tmp);
WaitOnPlaces(places);
auto sum_op = framework::OpRegistry::CreateOp(
"sum", {{"X", {s, tmp_name}}}, {{"Out", {s}}},
framework::AttributeMap{{"use_mkldnn", {false}}});
VLOG(10) << sum_op->DebugStringEx(sub_scopes[0]);
sum_op->Run(*sub_scopes[0], places[0]);
WaitOnPlace(places[0]);
}
CopyOrShare(*sub_scopes[0]->FindVar(s), place, scope.FindVar(s));
}
WaitOnPlaces(places);
}
};
std::ostream &operator<<(std::ostream &sout,
const std::vector<std::string> &strs) {
std::copy(strs.begin(), strs.end(),
std::ostream_iterator<std::string>(sout, ","));
return sout;
}
class ParallelDoGradOpDescMaker : public framework::SingleGradOpDescMaker {
public:
using framework::SingleGradOpDescMaker::SingleGradOpDescMaker;
protected:
virtual std::unique_ptr<framework::OpDesc> Apply() const {
auto *grad = new framework::OpDesc();
grad->SetType("parallel_do_grad");
for (auto &input_param : this->InputNames()) {
VLOG(3) << input_param;
grad->SetInput(input_param, this->Input(input_param));
if (input_param != kPlaces) {
grad->SetOutput(framework::GradVarName(input_param),
this->InputGrad(input_param, false));
}
}
auto *g_block = this->grad_block_[0];
// All variable name that needed by gradient operators
std::unordered_set<std::string> all_inputs_in_grad_blocks;
for (size_t i = 0; i < g_block->OpSize(); ++i) {
auto *op = g_block->Op(i);
for (auto &var_name : op->InputArgumentNames()) {
all_inputs_in_grad_blocks.insert(var_name);
}
}
for (auto &output_param : this->OutputNames()) {
if (output_param == kParallelScopes) {
grad->SetInput(output_param, this->Output(output_param));
grad->SetInput(framework::GradVarName(output_param),
this->Output(output_param));
} else {
grad->SetInput(output_param, this->Output(output_param));
std::vector<std::string> og_names;
for (auto &og_name : this->OutputGrad(output_param)) {
if (all_inputs_in_grad_blocks.count(og_name) != 0) {
// there are some gradient operators who need the OG. So make this
// OG as an input of parallel.do
og_names.push_back(og_name);
}
// else, there is no operator who need the OG. Do not use this OG as
// an input
}
grad->SetInput(framework::GradVarName(output_param), og_names);
}
}
grad->SetInput("Communicator", {"nccl_com__do_not_change_"});
grad->SetAttrMap(this->Attrs());
grad->SetBlockAttr(kParallelBlock, grad_block_[0]);
return std::unique_ptr<framework::OpDesc>(grad);
}
};
class ParallelDoGradOpShapeInference : public framework::InferShapeBase {
public:
void operator()(framework::InferShapeContext *ctx) const override {
PADDLE_ENFORCE(ctx->HasInputs(kParameters));
PADDLE_ENFORCE(ctx->HasInputs(kInputs));
PADDLE_ENFORCE(ctx->HasInputs(kOutputs));
ctx->SetOutputsDim(framework::GradVarName(kParameters),
ctx->GetInputsDim(kParameters));
auto i_dims = ctx->GetInputsDim(kInputs);
auto ig_names = ctx->Outputs(framework::GradVarName(kInputs));
for (size_t i = 0; i < ig_names.size(); ++i) {
auto &ig_name = ig_names[i];
if (ig_name == framework::kEmptyVarName) {
continue;
}
ctx->SetDims({ig_name}, {i_dims[i]});
}
auto p_dims = ctx->GetInputsDim(kParameters);
auto pg_names = ctx->Outputs(framework::GradVarName(kParameters));
for (size_t i = 0; i < pg_names.size(); ++i) {
auto &pg_name = pg_names[i];
if (pg_name == framework::kEmptyVarName) {
continue;
}
ctx->SetDims({pg_name}, {p_dims[i]});
}
}
};
class ParallelDoGradOpVarTypeInference : public framework::VarTypeInference {
public:
void operator()(const framework::OpDesc &op_desc,
framework::BlockDesc *block) const override {
framework::BlockDesc *sub_block =
boost::get<framework::BlockDesc *>(op_desc.GetAttr(kParallelBlock));
for (auto &out_vars : op_desc.Outputs()) {
for (auto &out_var : out_vars.second) {
auto &var = block->FindRecursiveOrCreateVar(out_var);
auto sub_var = sub_block->FindRecursiveOrCreateVar(out_var);
if (sub_var.GetType() != var.GetType()) {
var.SetType(sub_var.GetType());
}
}
}
}
};
} // namespace operators
} // namespace paddle
REGISTER_OPERATOR(parallel_do, paddle::operators::ParallelDoOp,
paddle::operators::ParallelDoOpProtoMaker,
paddle::operators::ParallelDoGradOpDescMaker);
REGISTER_OPERATOR(parallel_do_grad, paddle::operators::ParallelDoGradOp,
paddle::operators::ParallelDoGradOpShapeInference,
paddle::operators::ParallelDoGradOpVarTypeInference);
...@@ -150,19 +150,27 @@ class HierarchicalSigmoidGradOpKernel : public framework::OpKernel<T> { ...@@ -150,19 +150,27 @@ class HierarchicalSigmoidGradOpKernel : public framework::OpKernel<T> {
label.data<int64_t>())); label.data<int64_t>()));
} }
auto& place = *ctx.template device_context<DeviceContext>().eigen_device(); // softrelu derivative
auto pre_out_mat = EigenMatrix<T>::From(pre_out);
auto pre_out_grad_mat = EigenMatrix<T>::From(pre_out_grad);
auto out_grad_mat = EigenMatrix<T>::From(out_grad);
Eigen::array<int, 2> bcast{1, static_cast<int>(pre_out_grad.dims()[1])}; auto blas = math::GetBlas<DeviceContext, T>(ctx);
// softrelu derivative auto* pre_out_grad_data = pre_out_grad.data<T>();
pre_out_grad_mat.device(place) = auto* pre_out_data = pre_out.data<T>();
static_cast<T>(1.0) - static_cast<T>(1.0) / pre_out_mat.exp(); auto n = pre_out.numel();
blas.VEXP(n, pre_out_data, pre_out_grad_data);
blas.VINV(n, pre_out_grad_data, pre_out_grad_data);
for (int64_t i = 0; i < n; ++i) {
pre_out_grad_data[i] = 1.0 - pre_out_grad_data[i];
}
bit_code->Sub(&pre_out_grad); // the gradient of clip(w * x + b) bit_code->Sub(&pre_out_grad); // the gradient of clip(w * x + b)
pre_out_grad_mat.device(place) = auto* out_grad_data = out_grad.data<T>();
pre_out_grad_mat * out_grad_mat.broadcast(bcast);
int64_t dim0 = pre_out_grad.dims()[0];
int64_t dim1 = pre_out_grad.dims()[1];
for (int64_t i = 0; i < dim0; ++i) {
T tmp = out_grad_data[i];
blas.SCAL(dim1, tmp, pre_out_grad_data + i * dim1);
}
// TODO(guosheng): multiply pre_out_grad with subgradient of clipping to // TODO(guosheng): multiply pre_out_grad with subgradient of clipping to
// be consistent with the clipping in forward. // be consistent with the clipping in forward.
......
...@@ -124,8 +124,9 @@ REGISTER_OPERATOR(huber_loss, ops::HuberLossOp, ops::HuberLossOpMaker<float>, ...@@ -124,8 +124,9 @@ REGISTER_OPERATOR(huber_loss, ops::HuberLossOp, ops::HuberLossOpMaker<float>,
paddle::framework::DefaultGradOpDescMaker<true>); paddle::framework::DefaultGradOpDescMaker<true>);
REGISTER_OPERATOR(huber_loss_grad, ops::HuberLossGradOp); REGISTER_OPERATOR(huber_loss_grad, ops::HuberLossGradOp);
REGISTER_OP_CPU_KERNEL( REGISTER_OP_CPU_KERNEL(
huber_loss, huber_loss, ops::HuberLossKernel<paddle::platform::CPUDeviceContext, float>,
ops::HuberLossKernel<paddle::platform::CPUDeviceContext, float>); ops::HuberLossKernel<paddle::platform::CPUDeviceContext, double>);
REGISTER_OP_CPU_KERNEL( REGISTER_OP_CPU_KERNEL(
huber_loss_grad, huber_loss_grad,
ops::HuberLossGradKernel<paddle::platform::CPUDeviceContext, float>); ops::HuberLossGradKernel<paddle::platform::CPUDeviceContext, float>,
ops::HuberLossGradKernel<paddle::platform::CPUDeviceContext, double>);
...@@ -181,6 +181,9 @@ class Blas { ...@@ -181,6 +181,9 @@ class Blas {
const framework::Tensor& mat_b, const MatDescriptor& dim_b, const framework::Tensor& mat_b, const MatDescriptor& dim_b,
T alpha, framework::Tensor* mat_out, T beta) const; T alpha, framework::Tensor* mat_out, T beta) const;
template <typename T>
void VINV(int n, const T* a, T* y) const;
private: private:
const DeviceContext& context_; const DeviceContext& context_;
}; };
...@@ -282,6 +285,11 @@ class BlasT : private Blas<DeviceContext> { ...@@ -282,6 +285,11 @@ class BlasT : private Blas<DeviceContext> {
Base()->template BatchedGEMM<T>(args...); Base()->template BatchedGEMM<T>(args...);
} }
template <typename... ARGS>
void VINV(ARGS... args) const {
Base()->template VINV<T>(args...);
}
private: private:
const Blas<DeviceContext>* Base() const { const Blas<DeviceContext>* Base() const {
return static_cast<const Blas<DeviceContext>*>(this); return static_cast<const Blas<DeviceContext>*>(this);
......
...@@ -118,6 +118,11 @@ struct CBlas<float> { ...@@ -118,6 +118,11 @@ struct CBlas<float> {
static void VPOW(ARGS... args) { static void VPOW(ARGS... args) {
platform::dynload::vsPowx(args...); platform::dynload::vsPowx(args...);
} }
template <typename... ARGS>
static void VINV(ARGS... args) {
platform::dynload::vsInv(args...);
}
}; };
template <> template <>
...@@ -213,6 +218,11 @@ struct CBlas<double> { ...@@ -213,6 +218,11 @@ struct CBlas<double> {
static void VPOW(ARGS... args) { static void VPOW(ARGS... args) {
platform::dynload::vdPowx(args...); platform::dynload::vdPowx(args...);
} }
template <typename... ARGS>
static void VINV(ARGS... args) {
platform::dynload::vdInv(args...);
}
}; };
#else #else
...@@ -603,6 +613,17 @@ void Blas<DeviceContext>::MatMul(const framework::Tensor &mat_a, ...@@ -603,6 +613,17 @@ void Blas<DeviceContext>::MatMul(const framework::Tensor &mat_a,
dim_a.stride_, dim_b.stride_); dim_a.stride_, dim_b.stride_);
} }
} }
template <typename DeviceContext>
template <typename T>
void Blas<DeviceContext>::VINV(int n, const T *a, T *y) const {
#ifdef PADDLE_WITH_MKLML
CBlas<T>::VINV(n, a, y);
#else
for (int i = 0; i < n; ++i) {
y[i] = 1.0 / a[i];
}
#endif
}
} // namespace math } // namespace math
} // namespace operators } // namespace operators
......
...@@ -14,218 +14,380 @@ limitations under the License. */ ...@@ -14,218 +14,380 @@ limitations under the License. */
#include "paddle/fluid/operators/math/matrix_bit_code.h" #include "paddle/fluid/operators/math/matrix_bit_code.h"
#include <iostream> #include <iostream>
#include <map>
namespace paddle { namespace paddle {
namespace operators { namespace operators {
namespace math { namespace math {
template <typename T> template <typename T>
void MatrixBitCodeFunctor<T>::Add(const framework::Tensor& vec, struct MatrixBitCodeFunctorAdd : public boost::static_visitor<void> {
framework::Tensor* tmat) { const framework::Tensor &vec_;
size_t batch_size = tmat->dims()[0]; framework::Tensor *tmat_;
size_t width = tmat->dims()[1];
for (size_t i = 0; i < batch_size; ++i) { MatrixBitCodeFunctorAdd(const framework::Tensor &vec, framework::Tensor *tmat)
auto code = code_table_->get_code(i); : vec_(vec), tmat_(tmat) {}
int code_length = code->get_length();
for (int j = 0; j < code_length; ++j) { template <typename CodeTable>
size_t index = code->calc_index(j); void operator()(const CodeTable &code_table) {
tmat->data<T>()[i * width + j] += vec.data<T>()[index]; size_t batch_size = tmat_->dims()[0];
size_t width = tmat_->dims()[1];
auto *tmat_data = tmat_->data<T>();
auto *vec_data = vec_.data<T>();
for (size_t i = 0; i < batch_size; ++i) {
auto code = code_table.get_code(i);
int code_length = code.get_length();
for (int j = 0; j < code_length; ++j) {
size_t index = code.calc_index(j);
tmat_data[i * width + j] += vec_data[index];
}
} }
} }
};
template <typename T>
void MatrixBitCodeFunctor<T>::Add(const framework::Tensor &vec,
framework::Tensor *tmat) {
MatrixBitCodeFunctorAdd<T> func(vec, tmat);
code_table_.apply_visitor(func);
} }
template <typename T> template <typename T>
void MatrixBitCodeFunctor<T>::AddGrad(const framework::Tensor& tmat, struct MatrixBitCodeFunctorAddGrad : public boost::static_visitor<void> {
framework::Tensor* vec) { const framework::Tensor &tmat_;
size_t batch_size = tmat.dims()[0]; framework::Tensor *vec_;
size_t width = tmat.dims()[1]; MatrixBitCodeFunctorAddGrad(const framework::Tensor &tmat,
for (size_t i = 0; i < batch_size; ++i) { framework::Tensor *vec)
auto code = code_table_->get_code(i); : tmat_(tmat), vec_(vec) {}
int code_length = code->get_length();
for (int j = 0; j < code_length; ++j) { template <typename CodeTable>
size_t index = code->calc_index(j); void operator()(const CodeTable &table) {
vec->data<T>()[index] += tmat.data<T>()[i * width + j]; size_t batch_size = tmat_.dims()[0];
size_t width = tmat_.dims()[1];
auto *vec_data = vec_->data<T>();
auto *tmat_data = tmat_.data<T>();
for (size_t i = 0; i < batch_size; ++i) {
auto code = table.get_code(i);
int code_length = code.get_length();
for (int j = 0; j < code_length; ++j) {
size_t index = code.calc_index(j);
vec_data[index] += tmat_data[i * width + j];
}
} }
} }
};
template <typename T>
void MatrixBitCodeFunctor<T>::AddGrad(const framework::Tensor &tmat,
framework::Tensor *vec) {
MatrixBitCodeFunctorAddGrad<T> func(tmat, vec);
code_table_.apply_visitor(func);
} }
template <typename T> template <typename T>
void MatrixBitCodeFunctor<T>::AddGrad(const framework::Tensor& tmat, struct MatrixBitCodeFunctorSelectedRowsAddGrad
framework::SelectedRows* vec) { : public boost::static_visitor<void> {
size_t batch_size = tmat.dims()[0]; const framework::Tensor &tmat_;
size_t width = tmat.dims()[1]; framework::SelectedRows *vec_;
for (size_t i = 0; i < batch_size; ++i) {
auto code = code_table_->get_code(i); MatrixBitCodeFunctorSelectedRowsAddGrad(const framework::Tensor &tmat,
int code_length = code->get_length(); framework::SelectedRows *vec)
for (int j = 0; j < code_length; ++j) { : tmat_(tmat), vec_(vec) {}
size_t index = code->calc_index(j);
int64_t row_index = vec->GetIndexFromId(static_cast<int64_t>(index)); template <typename CodeTable>
vec->mutable_value()->data<T>()[row_index] += void operator()(const CodeTable &code_table) {
tmat.data<T>()[i * width + j]; size_t batch_size = tmat_.dims()[0];
size_t width = tmat_.dims()[1];
auto *vec_data = vec_->mutable_value()->template data<T>();
auto *tmat_data = tmat_.data<T>();
for (size_t i = 0; i < batch_size; ++i) {
auto code = code_table.get_code(i);
int code_length = code.get_length();
for (int j = 0; j < code_length; ++j) {
size_t index = code.calc_index(j);
int64_t row_index = vec_->GetIndexFromId(static_cast<int64_t>(index));
vec_data[row_index] += tmat_data[i * width + j];
}
} }
} }
};
template <typename T>
void MatrixBitCodeFunctor<T>::AddGrad(const framework::Tensor &tmat,
framework::SelectedRows *vec) {
MatrixBitCodeFunctorSelectedRowsAddGrad<T> func(tmat, vec);
code_table_.apply_visitor(func);
} }
template <typename T> template <typename T>
void MatrixBitCodeFunctor<T>::Sum(const framework::Tensor& tmat, struct MatrixBitCodeFunctorSum : public boost::static_visitor<void> {
framework::Tensor* sum, T scale_sum) { const framework::Tensor &tmat_;
size_t num_samples = tmat.dims()[0]; framework::Tensor *sum_;
size_t o_width = tmat.dims()[1]; T scale_sum_;
for (size_t i = 0; i < num_samples; ++i) {
T sm = static_cast<T>(0.0); MatrixBitCodeFunctorSum(const framework::Tensor &tmat, framework::Tensor *sum,
auto code = code_table_->get_code(i); T scale_sum)
int code_length = code->get_length(); : tmat_(tmat), sum_(sum), scale_sum_(scale_sum) {}
for (int j = 0; j < code_length; ++j) {
if (code->calc_bit(j)) { template <typename CodeTable>
// calc_bit starts from right most bit, while data in tmat[i] is in the void operator()(const CodeTable &code_table) {
// reverse order. size_t num_samples = tmat_.dims()[0];
sm += tmat.data<T>()[i * o_width + j]; size_t o_width = tmat_.dims()[1];
auto *tmat_data = tmat_.data<T>();
auto *sum_data = sum_->data<T>();
for (size_t i = 0; i < num_samples; ++i) {
T sm = static_cast<T>(0.0);
auto code = code_table.get_code(i);
int code_length = code.get_length();
for (int j = 0; j < code_length; ++j) {
if (code.calc_bit(j)) {
// calc_bit starts from right most bit, while data in tmat[i] is in
// the
// reverse order.
sm += tmat_data[i * o_width + j];
}
} }
sum_data[i] = scale_sum_ * sm;
} }
sum->data<T>()[i] = scale_sum * sm;
} }
};
template <typename T>
void MatrixBitCodeFunctor<T>::Sum(const framework::Tensor &tmat,
framework::Tensor *sum, T scale_sum) {
MatrixBitCodeFunctorSum<T> func(tmat, sum, scale_sum);
code_table_.apply_visitor(func);
} }
template <typename T> template <typename T>
void MatrixBitCodeFunctor<T>::Mul(framework::Tensor* tmat, struct MatrixBitCodeFunctorMul : public boost::static_visitor<void> {
const framework::Tensor& weight, framework::Tensor *tmat_;
const framework::Tensor& input) { const framework::Tensor &weight_;
auto blas = const framework::Tensor &input_;
GetBlas<platform::CPUDeviceContext, T>(platform::CPUDeviceContext());
size_t num_samples = tmat->dims()[0]; MatrixBitCodeFunctorMul(framework::Tensor *tmat,
size_t tmat_width = tmat->dims()[1]; const framework::Tensor &weight,
size_t input_width = input.dims()[1]; const framework::Tensor &input)
size_t weight_width = weight.dims()[1]; : tmat_(tmat), weight_(weight), input_(input) {}
auto tmat_value = tmat->data<T>();
auto weight_value = weight.data<T>(); template <typename CodeTable>
auto input_value = input.data<T>(); void operator()(const CodeTable &code_table) {
for (size_t i = 0; i < num_samples; ++i) { auto blas =
auto code = code_table_->get_code(i); GetBlas<platform::CPUDeviceContext, T>(platform::CPUDeviceContext());
int code_length = code->get_length(); size_t num_samples = tmat_->dims()[0];
const T* input_row = input_value + input_width * i; size_t tmat_width = tmat_->dims()[1];
for (int j = 0; j < code_length; ++j) { size_t input_width = input_.dims()[1];
size_t index = code->calc_index(j); size_t weight_width = weight_.dims()[1];
const T* weight_row = weight_value + weight_width * index; auto tmat_value = tmat_->data<T>();
T sum = static_cast<T>(0.0); auto weight_value = weight_.data<T>();
sum = blas.DOT(input_width, weight_row, input_row); auto input_value = input_.data<T>();
tmat_value[i * tmat_width + j] += sum; for (size_t i = 0; i < num_samples; ++i) {
auto code = code_table.get_code(i);
int code_length = code.get_length();
const T *input_row = input_value + input_width * i;
for (int j = 0; j < code_length; ++j) {
size_t index = code.calc_index(j);
const T *weight_row = weight_value + weight_width * index;
T sum = blas.DOT(input_width, weight_row, input_row);
tmat_value[i * tmat_width + j] += sum;
}
} }
} }
};
template <typename T>
void MatrixBitCodeFunctor<T>::Mul(framework::Tensor *tmat,
const framework::Tensor &weight,
const framework::Tensor &input) {
MatrixBitCodeFunctorMul<T> func(tmat, weight, input);
code_table_.apply_visitor(func);
} }
template <typename T, size_t N>
class ReservedVector : public std::vector<T> {
public:
ReservedVector() { this->reserve(N); }
};
template <typename T> template <typename T>
void MatrixBitCodeFunctor<T>::MulGradWeight(const framework::Tensor& tmat, struct MatrixBitCodeFunctorMulGradWeight : public boost::static_visitor<void> {
framework::Tensor* weight, const framework::Tensor &tmat_;
const framework::Tensor& input) { framework::Tensor *weight_;
auto blas = const framework::Tensor &input_;
GetBlas<platform::CPUDeviceContext, T>(platform::CPUDeviceContext()); MatrixBitCodeFunctorMulGradWeight(const framework::Tensor &tmat,
size_t num_samples = tmat.dims()[0]; framework::Tensor *weight,
size_t input_width = input.dims()[1]; const framework::Tensor &input)
size_t tmat_width = tmat.dims()[1]; : tmat_(tmat), weight_(weight), input_(input) {}
size_t weight_width = weight->dims()[1]; template <typename CodeTable>
auto tmat_value = tmat.data<T>(); void operator()(const CodeTable &code_table) {
auto weight_value = weight->data<T>(); auto blas =
auto input_value = input.data<T>(); GetBlas<platform::CPUDeviceContext, T>(platform::CPUDeviceContext());
size_t num_samples = tmat_.dims()[0];
std::unordered_map<int, std::vector<std::pair<T, const T*>>> ops; size_t input_width = input_.dims()[1];
size_t tmat_width = tmat_.dims()[1];
for (size_t i = 0; i < num_samples; ++i) { size_t weight_width = weight_->dims()[1];
auto code = code_table_->get_code(i); auto tmat_value = tmat_.data<T>();
int code_length = code->get_length(); auto weight_value = weight_->data<T>();
const T* input_value_row = input_value + input_width * i; auto input_value = input_.data<T>();
const T* tmat_row = tmat_value + i * tmat_width;
for (int j = 0; j < code_length; ++j) { std::map<int, ReservedVector<std::pair<T, const T *>, 8u>> ops;
ops[code->calc_index(j)].emplace_back(tmat_row[j], input_value_row); for (size_t i = 0; i < num_samples; ++i) {
auto code = code_table.get_code(i);
int code_length = code.get_length();
const T *input_value_row = input_value + input_width * i;
const T *tmat_row = tmat_value + i * tmat_width;
for (int j = 0; j < code_length; ++j) {
ops[code.calc_index(j)].emplace_back(tmat_row[j], input_value_row);
}
} }
} for (auto &op : ops) {
for (auto& op : ops) { auto &op_in_row = op.second;
auto& op_in_row = op.second; for (auto &pair : op_in_row) {
for (auto& pair : op_in_row) { auto &scale = pair.first;
auto& scale = pair.first; auto *input_row = pair.second;
auto* input_row = pair.second; T *weight_row = weight_value + op.first * weight_width;
T* weight_row = weight_value + op.first * weight_width; blas.AXPY(input_width, scale, input_row, weight_row);
blas.AXPY(input_width, scale, input_row, weight_row); }
} }
} }
};
template <typename T>
void MatrixBitCodeFunctor<T>::MulGradWeight(const framework::Tensor &tmat,
framework::Tensor *weight,
const framework::Tensor &input) {
MatrixBitCodeFunctorMulGradWeight<T> func(tmat, weight, input);
code_table_.apply_visitor(func);
} }
template <typename T> template <typename T>
void MatrixBitCodeFunctor<T>::MulGradWeight(const framework::Tensor& tmat, struct MatrixBitCodeFunctorMulGradWeightSR
framework::SelectedRows* weight, : public boost::static_visitor<void> {
const framework::Tensor& input) { const framework::Tensor &tmat_;
auto blas = framework::SelectedRows *weight_;
GetBlas<platform::CPUDeviceContext, T>(platform::CPUDeviceContext()); const framework::Tensor &input_;
size_t num_samples = tmat.dims()[0];
size_t input_width = input.dims()[1]; MatrixBitCodeFunctorMulGradWeightSR(const framework::Tensor &tmat,
size_t tmat_width = tmat.dims()[1]; framework::SelectedRows *weight,
size_t weight_width = weight->value().dims()[1]; const framework::Tensor &input)
auto tmat_value = tmat.data<T>(); : tmat_(tmat), weight_(weight), input_(input) {}
auto weight_value = weight->mutable_value()->data<T>();
auto input_value = input.data<T>(); template <typename CodeTable>
void operator()(const CodeTable &code_table) {
std::unordered_map<int, std::vector<std::pair<T, const T*>>> ops; auto blas =
ops.reserve(weight->rows().size()); GetBlas<platform::CPUDeviceContext, T>(platform::CPUDeviceContext());
size_t num_samples = tmat_.dims()[0];
for (size_t i = 0; i < num_samples; ++i) { size_t input_width = input_.dims()[1];
auto code = code_table_->get_code(i); size_t tmat_width = tmat_.dims()[1];
int code_length = code->get_length(); size_t weight_width = weight_->value().dims()[1];
const T* input_value_row = input_value + input_width * i; auto tmat_value = tmat_.data<T>();
const T* tmat_row = tmat_value + i * tmat_width; auto weight_value = weight_->mutable_value()->data<T>();
for (int j = 0; j < code_length; ++j) { auto input_value = input_.data<T>();
ops[code->calc_index(j)].emplace_back(tmat_row[j], input_value_row);
std::unordered_map<int, std::vector<std::pair<T, const T *>>> ops;
ops.reserve(weight_->rows().size());
for (size_t i = 0; i < num_samples; ++i) {
auto code = code_table.get_code(i);
int code_length = code.get_length();
const T *input_value_row = input_value + input_width * i;
const T *tmat_row = tmat_value + i * tmat_width;
for (int j = 0; j < code_length; ++j) {
ops[code.calc_index(j)].emplace_back(tmat_row[j], input_value_row);
}
} }
}
for (auto& row : weight->rows()) { for (auto &row : weight_->rows()) {
auto& op_in_row = ops[row]; auto &op_in_row = ops[row];
for (auto& pair : op_in_row) { for (auto &pair : op_in_row) {
auto& scale = pair.first; auto &scale = pair.first;
auto* input_row = pair.second; auto *input_row = pair.second;
blas.AXPY(input_width, scale, input_row, weight_value); blas.AXPY(input_width, scale, input_row, weight_value);
}
weight_value += weight_width;
} }
weight_value += weight_width;
} }
};
template <typename T>
void MatrixBitCodeFunctor<T>::MulGradWeight(const framework::Tensor &tmat,
framework::SelectedRows *weight,
const framework::Tensor &input) {
MatrixBitCodeFunctorMulGradWeightSR<T> func(tmat, weight, input);
code_table_.apply_visitor(func);
} }
template <typename T> template <typename T>
void MatrixBitCodeFunctor<T>::MulGradError(const framework::Tensor& tmat, struct MatrixBitCodeFunctorMulGradError : public boost::static_visitor<void> {
const framework::Tensor& weight, const framework::Tensor &tmat_;
framework::Tensor* input) { const framework::Tensor &weight_;
size_t num_samples = tmat.dims()[0]; framework::Tensor *input_;
size_t tmat_width = tmat.dims()[1];
size_t input_width = input->dims()[1]; MatrixBitCodeFunctorMulGradError(const framework::Tensor &tmat,
size_t weight_width = weight.dims()[1]; const framework::Tensor &weight,
auto tmat_value = tmat.data<T>(); framework::Tensor *input)
auto weight_value = weight.data<T>(); : tmat_(tmat), weight_(weight), input_(input) {}
auto input_value = input->data<T>(); template <typename CodeTable>
void operator()(const CodeTable &code_table) {
for (size_t i = 0; i < num_samples; ++i) { size_t num_samples = tmat_.dims()[0];
auto code = code_table_->get_code(i); size_t tmat_width = tmat_.dims()[1];
int code_length = code->get_length(); size_t input_width = input_->dims()[1];
for (int j = 0; j < code_length; ++j) { size_t weight_width = weight_.dims()[1];
size_t index = code->calc_index(j); auto tmat_value = tmat_.data<T>();
auto weight_value = weight_.data<T>();
for (size_t k = 0; k < input_width; ++k) { auto input_value = input_->data<T>();
input_value[input_width * i + k] +=
tmat_value[i * tmat_width + j] * for (size_t i = 0; i < num_samples; ++i) {
weight_value[weight_width * index + k]; auto code = code_table.get_code(i);
int code_length = code.get_length();
for (int j = 0; j < code_length; ++j) {
size_t index = code.calc_index(j);
for (size_t k = 0; k < input_width; ++k) {
input_value[input_width * i + k] +=
tmat_value[i * tmat_width + j] *
weight_value[weight_width * index + k];
}
} }
} }
} }
};
template <typename T>
void MatrixBitCodeFunctor<T>::MulGradError(const framework::Tensor &tmat,
const framework::Tensor &weight,
framework::Tensor *input) {
MatrixBitCodeFunctorMulGradError<T> func(tmat, weight, input);
code_table_.apply_visitor(func);
} }
template <typename T> template <typename T>
void MatrixBitCodeFunctor<T>::Sub(framework::Tensor* tmat) { struct MatrixBitCodeFunctorSub : public boost::static_visitor<void> {
size_t num_samples = tmat->dims()[0]; framework::Tensor *tmat_;
size_t o_width = tmat->dims()[1];
for (size_t i = 0; i < num_samples; ++i) { explicit MatrixBitCodeFunctorSub(framework::Tensor *tmat) : tmat_(tmat) {}
auto code = code_table_->get_code(i);
int code_length = code->get_length(); template <typename CodeTable>
for (int j = 0; j < code_length; ++j) { void operator()(const CodeTable &code_table) {
if (code->calc_bit(j)) { size_t num_samples = tmat_->dims()[0];
tmat->data<T>()[i * o_width + j] -= 1; size_t o_width = tmat_->dims()[1];
auto *tmat_data = tmat_->data<T>();
for (size_t i = 0; i < num_samples; ++i) {
auto code = code_table.get_code(i);
int code_length = code.get_length();
for (int j = 0; j < code_length; ++j) {
if (code.calc_bit(j)) {
tmat_data[i * o_width + j] -= 1;
}
} }
} }
} }
};
template <typename T>
void MatrixBitCodeFunctor<T>::Sub(framework::Tensor *tmat) {
MatrixBitCodeFunctorSub<T> func(tmat);
code_table_.apply_visitor(func);
} }
template class MatrixBitCodeFunctor<float>; template class MatrixBitCodeFunctor<float>;
......
...@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and ...@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and
limitations under the License. */ limitations under the License. */
#pragma once #pragma once
#include <map>
#include <unordered_map> #include <unordered_map>
#include <utility> #include <utility>
#include <vector> #include <vector>
...@@ -22,6 +23,7 @@ limitations under the License. */ ...@@ -22,6 +23,7 @@ limitations under the License. */
#include "paddle/fluid/framework/tensor.h" #include "paddle/fluid/framework/tensor.h"
#include "paddle/fluid/operators/math/blas.h" #include "paddle/fluid/operators/math/blas.h"
#include "paddle/fluid/platform/device_context.h" #include "paddle/fluid/platform/device_context.h"
#include "paddle/fluid/platform/variant.h"
#if defined(_WIN32) #if defined(_WIN32)
#include <intrin.h> #include <intrin.h>
...@@ -98,24 +100,7 @@ inline int clz(const T& value) { ...@@ -98,24 +100,7 @@ inline int clz(const T& value) {
inline size_t FindLastSet(size_t x) { return sizeof(size_t) * 8 - clz(x); } inline size_t FindLastSet(size_t x) { return sizeof(size_t) * 8 - clz(x); }
#endif // !_WIN32 #endif // !_WIN32
// set a code interface to create multiple code class SimpleCode {
class Code {
public:
virtual ~Code() {}
virtual size_t calc_index(int bit) const = 0;
virtual bool calc_bit(int bit) const = 0;
virtual int get_length() const = 0;
};
// set a CodeTable interface to create multiple code table
class CodeTable {
public:
virtual std::unique_ptr<Code> get_code(int64_t code) const = 0;
virtual size_t size() const = 0;
virtual int get_max_code_length() const = 0;
virtual ~CodeTable() {}
};
class SimpleCode : public Code {
public: public:
SimpleCode(size_t code, size_t num_classes, const int64_t* ids) SimpleCode(size_t code, size_t num_classes, const int64_t* ids)
: c_(static_cast<size_t>(ids[code]) + num_classes) {} : c_(static_cast<size_t>(ids[code]) + num_classes) {}
...@@ -137,16 +122,16 @@ class SimpleCode : public Code { ...@@ -137,16 +122,16 @@ class SimpleCode : public Code {
}; };
template <typename T> template <typename T>
class CustomCode : public Code { class CustomCode {
public: public:
CustomCode(const framework::Tensor& ptable, const framework::Tensor& pcode, CustomCode(const framework::Tensor& ptable, const framework::Tensor& pcode,
const int64_t* ids, int index) const int64_t* ids, int index) {
: ids_(ids), index_(index) { seq_len_ = ptable.dims()[1];
ptable_ = ptable.Slice(index, index + 1); ptable_data_ = ptable.data<T>() + seq_len_ * index;
pcode_ = pcode.Slice(index, index + 1); pcode_data_ = pcode.data<T>() + seq_len_ * index;
} }
/** /**
* Here the id of root shoud be 1 rather than 0, thus the encoding of class c * Here the id of root should be 1 rather than 0, thus the encoding of class c
* is `c + num_classes` and all siblings can get the same weight indice using * is `c + num_classes` and all siblings can get the same weight indice using
* prefixes. * prefixes.
* Weight index is the prefixes of encoding, thus leave out the right most * Weight index is the prefixes of encoding, thus leave out the right most
...@@ -154,36 +139,37 @@ class CustomCode : public Code { ...@@ -154,36 +139,37 @@ class CustomCode : public Code {
* Binary classification path is the suffixes of encoding, thus leave out the * Binary classification path is the suffixes of encoding, thus leave out the
* left most bit in calc_bit. * left most bit in calc_bit.
*/ */
size_t calc_index(int bit) const { return ptable_.data<T>()[bit]; } size_t calc_index(int bit) const { return ptable_data_[bit]; }
bool calc_bit(int bit) const { return pcode_.data<T>()[bit]; } bool calc_bit(int bit) const { return pcode_data_[bit]; }
int get_length() const {
int length = 0;
for (int i = 0; i < static_cast<int>(ptable_.dims()[1]); i++) { // NOTE: this function is not thread-safe.
if (ptable_.data<T>()[i] >= 0) { int get_length() const {
length++; if (length_ < 0) {
} else { auto len = seq_len_;
return length; length_ =
} static_cast<int>(std::find_if(ptable_data_, ptable_data_ + len,
[](const T& val) { return val < 0; }) -
ptable_data_);
} }
return length; return length_;
} }
private: private:
framework::Tensor ptable_; int64_t seq_len_;
framework::Tensor pcode_; const T* ptable_data_;
const int64_t* ids_; const T* pcode_data_;
const int index_; mutable int length_{-1};
}; };
class SimpleCodeTable : public CodeTable { class SimpleCodeTable {
public: public:
SimpleCodeTable(size_t num_classes, const int64_t* ids) SimpleCodeTable(size_t num_classes, const int64_t* ids)
: num_classes_(num_classes), ids_(ids) {} : num_classes_(num_classes), ids_(ids) {}
std::unique_ptr<Code> get_code(int64_t code) const {
std::unique_ptr<Code> coder(new SimpleCode(code, num_classes_, ids_)); SimpleCode get_code(int64_t code) const {
return coder; return SimpleCode(code, num_classes_, ids_);
} }
size_t size() const { return num_classes_; } size_t size() const { return num_classes_; }
int get_max_code_length() const { return FindLastSet(num_classes_ - 1); } int get_max_code_length() const { return FindLastSet(num_classes_ - 1); }
...@@ -193,15 +179,14 @@ class SimpleCodeTable : public CodeTable { ...@@ -193,15 +179,14 @@ class SimpleCodeTable : public CodeTable {
}; };
template <typename T> template <typename T>
class CustomCodeTable : public CodeTable { class CustomCodeTable {
public: public:
CustomCodeTable(const framework::Tensor& ptable, CustomCodeTable(const framework::Tensor& ptable,
const framework::Tensor& pcode, const int64_t* ids) const framework::Tensor& pcode, const int64_t* ids)
: ptable_(ptable), pcode_(pcode), ids_(ids) {} : ptable_(ptable), pcode_(pcode), ids_(ids) {}
std::unique_ptr<Code> get_code(int64_t code) const { CustomCode<T> get_code(int64_t code) const {
std::unique_ptr<Code> coder(new CustomCode<T>(ptable_, pcode_, ids_, code)); return CustomCode<T>(ptable_, pcode_, ids_, code);
return coder;
} }
size_t size() const { return static_cast<size_t>(ptable_.dims()[1]); } size_t size() const { return static_cast<size_t>(ptable_.dims()[1]); }
...@@ -215,19 +200,21 @@ class CustomCodeTable : public CodeTable { ...@@ -215,19 +200,21 @@ class CustomCodeTable : public CodeTable {
const int64_t* ids_; const int64_t* ids_;
}; };
using CodeTable = boost::variant<SimpleCodeTable, CustomCodeTable<int64_t>>;
template <typename T> template <typename T>
class MatrixBitCodeFunctor { class MatrixBitCodeFunctor {
public: public:
MatrixBitCodeFunctor(size_t num_classes, const int64_t* ids) MatrixBitCodeFunctor(size_t num_classes, const int64_t* ids)
: num_classes_(num_classes), : num_classes_(num_classes),
ids_(ids), ids_(ids),
code_table_(new SimpleCodeTable(num_classes, ids)) {} code_table_(SimpleCodeTable(num_classes, ids)) {}
MatrixBitCodeFunctor(const framework::Tensor& ptable, MatrixBitCodeFunctor(const framework::Tensor& ptable,
const framework::Tensor& pcode, const int64_t* ids) const framework::Tensor& pcode, const int64_t* ids)
: num_classes_(static_cast<size_t>(ptable.dims()[1])), : num_classes_(static_cast<size_t>(ptable.dims()[1])),
ids_(ids), ids_(ids),
code_table_(new CustomCodeTable<int64_t>(ptable, pcode, ids)) {} code_table_(CustomCodeTable<int64_t>(ptable, pcode, ids)) {}
/* For j < code_length /* For j < code_length
tmat(i, j) += vec(0, index(i, j)) tmat(i, j) += vec(0, index(i, j))
*/ */
...@@ -277,7 +264,7 @@ class MatrixBitCodeFunctor { ...@@ -277,7 +264,7 @@ class MatrixBitCodeFunctor {
size_t num_classes_; size_t num_classes_;
const int64_t* ids_; const int64_t* ids_;
std::unique_ptr<CodeTable> code_table_; CodeTable code_table_;
}; };
} // namespace math } // namespace math
} // namespace operators } // namespace operators
......
...@@ -109,6 +109,11 @@ class AdamOpMaker : public framework::OpProtoAndCheckerMaker { ...@@ -109,6 +109,11 @@ class AdamOpMaker : public framework::OpProtoAndCheckerMaker {
"(float, default 1.0e-8) " "(float, default 1.0e-8) "
"Constant for numerical stability") "Constant for numerical stability")
.SetDefault(1.0e-8f); .SetDefault(1.0e-8f);
AddAttr<bool>(
"lazy_mode",
"(bool, default false) "
"only update the parameter that has gradient in sparse update")
.SetDefault(false);
AddComment(R"DOC( AddComment(R"DOC(
Adam Optimizer. Adam Optimizer.
......
...@@ -177,12 +177,13 @@ struct SparseAdamFunctor { ...@@ -177,12 +177,13 @@ struct SparseAdamFunctor {
const int64_t* rows_; const int64_t* rows_;
int64_t row_numel_; int64_t row_numel_;
int64_t row_count_; int64_t row_count_;
bool lazy_mode_;
SparseAdamFunctor(T beta1, T beta2, T epsilon, const T* beta1_pow, SparseAdamFunctor(T beta1, T beta2, T epsilon, const T* beta1_pow,
const T* beta2_pow, const T* mom1, T* mom1_out, const T* beta2_pow, const T* mom1, T* mom1_out,
const T* mom2, T* mom2_out, const T* lr, const T* grad, const T* mom2, T* mom2_out, const T* lr, const T* grad,
const T* param, T* param_out, const int64_t* rows, const T* param, T* param_out, const int64_t* rows,
int64_t row_numel, int64_t row_count) int64_t row_numel, int64_t row_count, bool lazy_mode)
: beta1_(beta1), : beta1_(beta1),
beta2_(beta2), beta2_(beta2),
epsilon_(epsilon), epsilon_(epsilon),
...@@ -198,13 +199,10 @@ struct SparseAdamFunctor { ...@@ -198,13 +199,10 @@ struct SparseAdamFunctor {
param_out_(param_out), param_out_(param_out),
rows_(rows), rows_(rows),
row_numel_(row_numel), row_numel_(row_numel),
row_count_(row_count) {} row_count_(row_count),
lazy_mode_(lazy_mode) {}
inline HOSTDEVICE void operator()(size_t i) const {
auto row_idx =
math::BinarySearch<int64_t>(rows_, row_count_, i / row_numel_);
T g = row_idx >= 0 ? grad_[row_idx * row_numel_ + i % row_numel_] : 0;
inline HOSTDEVICE void adam_update(size_t i, T g) const {
// The following code is the same as dense // The following code is the same as dense
T mom1 = moment1_[i]; T mom1 = moment1_[i];
T mom2 = moment2_[i]; T mom2 = moment2_[i];
...@@ -225,6 +223,17 @@ struct SparseAdamFunctor { ...@@ -225,6 +223,17 @@ struct SparseAdamFunctor {
moment2_out_[i] = mom2; moment2_out_[i] = mom2;
param_out_[i] = p; param_out_[i] = p;
} }
inline HOSTDEVICE void operator()(size_t i) const {
auto row_idx =
math::BinarySearch<int64_t>(rows_, row_count_, i / row_numel_);
if (lazy_mode_ && row_idx < 0) {
return;
} else {
T g = row_idx >= 0 ? grad_[row_idx * row_numel_ + i % row_numel_] : 0;
adam_update(i, g);
}
}
}; };
template <typename DeviceContext, typename T> template <typename DeviceContext, typename T>
...@@ -240,6 +249,7 @@ class AdamOpKernel : public framework::OpKernel<T> { ...@@ -240,6 +249,7 @@ class AdamOpKernel : public framework::OpKernel<T> {
using paddle::framework::LoDTensor; using paddle::framework::LoDTensor;
using paddle::operators::detail::Ref; using paddle::operators::detail::Ref;
bool lazy_mode = ctx.Attr<bool>("lazy_mode");
T beta1 = static_cast<T>(ctx.Attr<float>("beta1")); T beta1 = static_cast<T>(ctx.Attr<float>("beta1"));
T beta2 = static_cast<T>(ctx.Attr<float>("beta2")); T beta2 = static_cast<T>(ctx.Attr<float>("beta2"));
T epsilon = static_cast<T>(ctx.Attr<float>("epsilon")); T epsilon = static_cast<T>(ctx.Attr<float>("epsilon"));
...@@ -351,11 +361,23 @@ class AdamOpKernel : public framework::OpKernel<T> { ...@@ -351,11 +361,23 @@ class AdamOpKernel : public framework::OpKernel<T> {
mom2_out.template mutable_data<T>(ctx.GetPlace()), mom2_out.template mutable_data<T>(ctx.GetPlace()),
lr.template data<T>(), grad_data, param.template data<T>(), lr.template data<T>(), grad_data, param.template data<T>(),
param_out.template mutable_data<T>(ctx.GetPlace()), rows, row_numel, param_out.template mutable_data<T>(ctx.GetPlace()), rows, row_numel,
grad_merge.rows().size()); grad_merge.rows().size(), lazy_mode);
platform::ForRange<DeviceContext> for_range( VLOG(3) << "lazy_mode :" << lazy_mode;
static_cast<const DeviceContext&>(ctx.device_context()), if (lazy_mode && platform::is_cpu_place(ctx.GetPlace())) {
param.numel()); size_t row_count = grad_merge.rows().size();
for_range(functor); std::vector<int64_t> cpu_rows(grad_merge.rows());
for (size_t row_index = 0; row_index < row_count; ++row_index) {
for (size_t offset = 0; offset < row_numel; ++offset) {
size_t i = cpu_rows[row_index] * row_numel + offset;
functor.adam_update(i, grad_data[row_index * row_numel + offset]);
}
}
} else {
platform::ForRange<DeviceContext> for_range(
static_cast<const DeviceContext&>(ctx.device_context()),
param.numel());
for_range(functor);
}
} else { } else {
PADDLE_THROW("Variable type not supported by adam_op"); PADDLE_THROW("Variable type not supported by adam_op");
} }
......
op_library(tensorrt_engine_op DEPS tensorrt_engine tensorrt_converter) op_library(tensorrt_engine_op DEPS tensorrt_engine tensorrt_converter)
file(APPEND ${pybind_file} "USE_CUDA_ONLY_OP(tensorrt_engine);\n") file(APPEND ${pybind_file} "USE_NO_KERNEL_OP(tensorrt_engine);\n")
nv_test(test_tensorrt_engine_op SRCS tensorrt_engine_op_test.cc nv_test(test_tensorrt_engine_op SRCS tensorrt_engine_op_test.cc
DEPS tensorrt_engine_op DEPS tensorrt_engine_op
analysis) analysis)
...@@ -21,8 +21,6 @@ ...@@ -21,8 +21,6 @@
namespace paddle { namespace paddle {
DEFINE_int32(tensorrt_engine_batch_size, 1, "the batch_size of TensorRT");
namespace operators { namespace operators {
class TensorRTEngineOpMaker : public framework::OpProtoAndCheckerMaker { class TensorRTEngineOpMaker : public framework::OpProtoAndCheckerMaker {
...@@ -31,7 +29,6 @@ class TensorRTEngineOpMaker : public framework::OpProtoAndCheckerMaker { ...@@ -31,7 +29,6 @@ class TensorRTEngineOpMaker : public framework::OpProtoAndCheckerMaker {
AddInput("Xs", "A list of inputs.").AsDuplicable(); AddInput("Xs", "A list of inputs.").AsDuplicable();
AddOutput("Ys", "A list of outputs").AsDuplicable(); AddOutput("Ys", "A list of outputs").AsDuplicable();
AddAttr<std::string>("subgraph", "the subgraph."); AddAttr<std::string>("subgraph", "the subgraph.");
AddAttr<std::string>("engine_uniq_key", "unique key for the TRT engine.");
AddAttr<int>("max_batch_size", "the maximum batch size."); AddAttr<int>("max_batch_size", "the maximum batch size.");
AddAttr<int>("workspace_size", "the workspace size."); AddAttr<int>("workspace_size", "the workspace size.");
AddComment("TensorRT engine operator."); AddComment("TensorRT engine operator.");
...@@ -50,6 +47,6 @@ class TensorRTEngineInferVarType : public framework::VarTypeInference { ...@@ -50,6 +47,6 @@ class TensorRTEngineInferVarType : public framework::VarTypeInference {
namespace ops = paddle::operators; namespace ops = paddle::operators;
REGISTER_OPERATOR(tensorrt_engine, ops::TensorRTEngineOp, REGISTER_OPERATOR(tensorrt_engine, ops::TensorRTEngineOp,
ops::TensorRTEngineOpMaker, ops::TensorRTEngineOpMaker); ops::TensorRTEngineOpMaker);
#endif // PADDLE_WITH_CUDA #endif // PADDLE_WITH_CUDA
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License. */
#include "paddle/fluid/operators/tensorrt/tensorrt_engine_op.h"
namespace ops = paddle::operators;
REGISTER_OP_CUDA_KERNEL(
tensorrt_engine,
ops::TensorRTEngineKernel<paddle::platform::CUDADeviceContext, float>,
ops::TensorRTEngineKernel<paddle::platform::CUDADeviceContext, double>,
ops::TensorRTEngineKernel<paddle::platform::CUDADeviceContext, int>,
ops::TensorRTEngineKernel<paddle::platform::CUDADeviceContext, int64_t>);
...@@ -27,8 +27,6 @@ ...@@ -27,8 +27,6 @@
namespace paddle { namespace paddle {
DECLARE_int32(tensorrt_engine_batch_size);
namespace operators { namespace operators {
using FluidDT = framework::proto::VarType_Type; using FluidDT = framework::proto::VarType_Type;
...@@ -49,7 +47,7 @@ TRT_DT FluidDataType2TRT(FluidDT type) { ...@@ -49,7 +47,7 @@ TRT_DT FluidDataType2TRT(FluidDT type) {
return TRT_DT::kINT32; return TRT_DT::kINT32;
} }
nvinfer1::Dims Vec2TRT_Dims(const std::vector<int64_t>& shape) { nvinfer1::Dims Vec2TRT_Dims(const std::vector<int64_t> &shape) {
PADDLE_ENFORCE_GT(shape.size(), 1UL, PADDLE_ENFORCE_GT(shape.size(), 1UL,
"TensorRT' tensor input requires at least 2 dimensions"); "TensorRT' tensor input requires at least 2 dimensions");
PADDLE_ENFORCE_LE(shape.size(), 4UL, PADDLE_ENFORCE_LE(shape.size(), 4UL,
...@@ -63,128 +61,119 @@ nvinfer1::Dims Vec2TRT_Dims(const std::vector<int64_t>& shape) { ...@@ -63,128 +61,119 @@ nvinfer1::Dims Vec2TRT_Dims(const std::vector<int64_t>& shape) {
} // namespace // NOLINT } // namespace // NOLINT
using inference::Singleton; using inference::Singleton;
using inference::tensorrt::TRT_EngineManager; using inference::tensorrt::TensorRTEngine;
class TensorRTEngineOp : public framework::OperatorBase {
private:
std::vector<std::string> input_names_;
std::unordered_set<std::string> param_names_;
mutable std::unique_ptr<TensorRTEngine> trt_engine_;
int max_batch_size_;
int workspace_size_;
class TensorRTEngineOp : public framework::OperatorWithKernel {
public: public:
using framework::OperatorWithKernel::OperatorWithKernel; TensorRTEngineOp(const std::string &type,
const framework::VariableNameMap &inputs,
const framework::VariableNameMap &outputs,
const framework::AttributeMap &attrs)
: framework::OperatorBase(type, inputs, outputs, attrs) {
input_names_ = Inputs("Xs");
max_batch_size_ = Attr<int>("max_batch_size");
workspace_size_ = Attr<int>("workspace_size");
auto params = Attr<std::vector<std::string>>("parameters");
for (const auto &param : params) {
param_names_.insert(param);
}
}
protected: protected:
void InferShape(framework::InferShapeContext* ctx) const override {} void RunImpl(const framework::Scope &scope,
const platform::Place &dev_place) const override {
framework::OpKernelType GetExpectedKernelType( RunTrt(scope, dev_place);
const framework::ExecutionContext& ctx) const override {
auto input0 = ctx.Inputs("Xs").front();
framework::OpKernelType kt = framework::OpKernelType(
ctx.scope().FindVar(input0)->GetMutable<framework::LoDTensor>()->type(),
ctx.GetPlace());
return kt;
} }
};
template <typename DeviceContext, typename T> void RunTrt(const framework::Scope &scope,
class TensorRTEngineKernel : public framework::OpKernel<T> { const platform::Place &dev_place) const {
public: int runtime_batch = 1;
void Compute(const framework::ExecutionContext& context) const override { if (trt_engine_.get() == nullptr) {
auto engine_name = context.Attr<std::string>("engine_uniq_key"); trt_engine_.reset(new TensorRTEngine(
int max_batch_size = context.Attr<int>("max_batch_size"); max_batch_size_, workspace_size_, nullptr,
if (!Singleton<TRT_EngineManager>::Global().HasEngine(engine_name)) { boost::get<platform::CUDAPlace>(dev_place).device));
Prepare(context); Prepare(scope, dev_place, trt_engine_.get());
} }
auto* engine = Singleton<TRT_EngineManager>::Global().Get(engine_name);
auto input_names = context.op().Inputs("Xs"); auto *engine = trt_engine_.get();
PADDLE_ENFORCE(!input_names.empty(), "should pass more than one inputs"); PADDLE_ENFORCE(!input_names_.empty(), "should pass more than one inputs");
PADDLE_ENFORCE_LE(FLAGS_tensorrt_engine_batch_size, max_batch_size);
std::vector<std::string> output_maps = std::vector<std::string> output_maps =
context.Attr<std::vector<std::string>>("output_name_mapping"); Attr<std::vector<std::string>>("output_name_mapping");
auto params = context.Attr<std::vector<std::string>>("parameters");
std::unordered_set<std::string> parameters;
for (const auto& param : params) {
parameters.insert(param);
}
// Convert input tensor from fluid to engine. // Convert input tensor from fluid to engine.
for (const auto& x : context.Inputs("Xs")) { for (const auto &x : Inputs("Xs")) {
if (parameters.count(x)) continue; if (param_names_.count(x)) continue;
// convert input and copy to TRT engine's buffer // convert input and copy to TRT engine's buffer
auto& t = inference::analysis::GetFromScope<framework::LoDTensor>( auto &t =
context.scope(), x); inference::analysis::GetFromScope<framework::LoDTensor>(scope, x);
auto t_shape = framework::vectorize(t.dims());
runtime_batch = t_shape[0];
if (platform::is_cpu_place(t.place())) { if (platform::is_cpu_place(t.place())) {
engine->SetInputFromCPU(x, static_cast<const void*>(t.data<void>()), engine->SetInputFromCPU(x, static_cast<const void *>(t.data<void>()),
t.memory_size()); t.memory_size());
} else { } else {
engine->SetInputFromGPU(x, static_cast<const void*>(t.data<void>()), engine->SetInputFromGPU(x, static_cast<const void *>(t.data<void>()),
t.memory_size()); t.memory_size());
} }
} }
PADDLE_ENFORCE_LE(runtime_batch, max_batch_size_);
// Execute the engine. // Execute the engine.
PADDLE_ENFORCE_GT(FLAGS_tensorrt_engine_batch_size, 0); engine->Execute(runtime_batch);
engine->Execute(FLAGS_tensorrt_engine_batch_size);
// Convert output tensor from engine to fluid // Convert output tensor from engine to fluid
int output_index = 0; int output_index = 0;
VLOG(4) << "TensorRT Engine Op Outputs:"; VLOG(4) << "TensorRT Engine Op Outputs:";
for (const auto& y : context.Outputs("Ys")) { for (const auto &y : Outputs("Ys")) {
VLOG(4) << y; VLOG(4) << y;
// convert output and copy to fluid. // convert output and copy to fluid.
nvinfer1::ITensor* trt_t = engine->GetITensor(output_maps[output_index]); nvinfer1::ITensor *trt_t = engine->GetITensor(output_maps[output_index]);
auto dims = trt_t->getDimensions(); auto dims = trt_t->getDimensions();
// Use the output ITensor's dims to reshape the Fluid Tensor. // Use the output ITensor's dims to reshape the Fluid Tensor.
// The ITensor doesn't contain the batch size dim. // The ITensor doesn't contain the batch size dim.
std::vector<int> ddim; std::vector<int> ddim;
ddim.push_back(FLAGS_tensorrt_engine_batch_size); ddim.push_back(runtime_batch);
for (int i = 0; i < dims.nbDims; i++) { for (int i = 0; i < dims.nbDims; i++) {
ddim.push_back(dims.d[i]); ddim.push_back(dims.d[i]);
} }
auto* fluid_v = context.scope().FindVar(y); auto *fluid_v = scope.FindVar(y);
PADDLE_ENFORCE_NOT_NULL(fluid_v, "no output variable called %s", y); PADDLE_ENFORCE_NOT_NULL(fluid_v, "no output variable called %s", y);
auto* fluid_t = fluid_v->GetMutable<framework::LoDTensor>(); auto *fluid_t = fluid_v->GetMutable<framework::LoDTensor>();
fluid_t->Resize(framework::make_ddim(ddim)); fluid_t->Resize(framework::make_ddim(ddim));
// TODO(Superjomn) find some way to determine which device to output the
// tensor.
// if (platform::is_cpu_place(fluid_t->place())) {
// TODO(Superjomn) change this float to dtype size. // TODO(Superjomn) change this float to dtype size.
auto size = inference::analysis::AccuDims(dims.d, dims.nbDims) * auto size =
FLAGS_tensorrt_engine_batch_size; inference::analysis::AccuDims(dims.d, dims.nbDims) * runtime_batch;
engine->GetOutputInGPU( engine->GetOutputInGPU(
output_maps[output_index], output_maps[output_index],
fluid_t->mutable_data<float>(platform::CUDAPlace( fluid_t->mutable_data<float>(platform::CUDAPlace(
boost::get<platform::CUDAPlace>(context.GetPlace()).device)), boost::get<platform::CUDAPlace>(dev_place).device)),
size * sizeof(float)); size * sizeof(float));
output_index += 1; output_index += 1;
} }
cudaStreamSynchronize(*engine->stream()); cudaStreamSynchronize(*engine->stream());
} }
protected: void Prepare(const framework::Scope &scope, const platform::Place &dev_place,
void Prepare(const framework::ExecutionContext& context) const { TensorRTEngine *engine) const {
VLOG(4) << "Prepare engine"; VLOG(4) << "Prepare engine";
// Get the ProgramDesc and pass to convert.
framework::proto::BlockDesc block_desc; framework::proto::BlockDesc block_desc;
block_desc.ParseFromString(context.Attr<std::string>("subgraph")); block_desc.ParseFromString(Attr<std::string>("subgraph"));
int max_batch_size = context.Attr<int>("max_batch_size");
int workspace_size = context.Attr<int>("workspace_size");
auto params = context.Attr<std::vector<std::string>>("parameters");
std::unordered_set<std::string> parameters;
for (const auto& param : params) {
parameters.insert(param);
}
std::vector<std::string> output_maps = std::vector<std::string> output_maps =
context.Attr<std::vector<std::string>>("output_name_mapping"); Attr<std::vector<std::string>>("output_name_mapping");
// TODO(Superjomn) replace this with a different stream
auto* engine = Singleton<TRT_EngineManager>::Global().Create(
max_batch_size, workspace_size, nullptr /*engine hold its own stream*/,
context.Attr<std::string>("engine_uniq_key"),
boost::get<platform::CUDAPlace>(context.GetPlace()).device);
engine->InitNetwork(); engine->InitNetwork();
...@@ -192,39 +181,33 @@ class TensorRTEngineKernel : public framework::OpKernel<T> { ...@@ -192,39 +181,33 @@ class TensorRTEngineKernel : public framework::OpKernel<T> {
VLOG(4) << "parsed var size " << block.AllVars().size(); VLOG(4) << "parsed var size " << block.AllVars().size();
// Add inputs // Add inputs
VLOG(4) << "declare inputs"; VLOG(4) << "declare inputs";
for (auto& input : context.Inputs("Xs")) { for (auto &input : Inputs("Xs")) {
if (parameters.count(input)) continue; if (param_names_.count(input)) continue;
VLOG(4) << "declare input " << input; VLOG(4) << "declare input " << input;
auto* var = block.FindVar(input);
auto &t =
inference::analysis::GetFromScope<framework::LoDTensor>(scope, input);
auto t_shape = framework::vectorize(t.dims());
auto *var = block.FindVar(input);
// TensorRT engine need to create parameters. The parameter's description // TensorRT engine need to create parameters. The parameter's description
// should be set in // should be set in
PADDLE_ENFORCE(var, "no variable called %s", input); PADDLE_ENFORCE(var, "no variable called %s", input);
PADDLE_ENFORCE_EQ(var->GetType(), FluidDT::VarType_Type_LOD_TENSOR, PADDLE_ENFORCE_EQ(var->GetType(), FluidDT::VarType_Type_LOD_TENSOR,
"TensorRT engine only takes LoDTensor as input"); "TensorRT engine only takes LoDTensor as input");
auto shape = var->GetShape();
// For the special batch_size placeholder -1, drop it and pass the real
// shape of data.
// TODO(Superjomn) fix this with batch broadcast, or it can't handle
// variational batch size.
if (shape[0] == -1) {
shape[0] = FLAGS_tensorrt_engine_batch_size;
}
engine->DeclareInput( engine->DeclareInput(
input, FluidDataType2TRT( input, FluidDataType2TRT(
var->Proto()->type().lod_tensor().tensor().data_type()), var->Proto()->type().lod_tensor().tensor().data_type()),
Vec2TRT_Dims(shape)); Vec2TRT_Dims(t_shape));
} }
inference::Singleton<inference::tensorrt::OpConverter>::Global() inference::Singleton<inference::tensorrt::OpConverter>::Global()
.ConvertBlock(block_desc, parameters, context.scope(), engine); .ConvertBlock(block_desc, param_names_, scope, engine);
// Add outputs // Add outputs
for (auto& output : output_maps) { for (auto &output : output_maps) {
if (!engine->HasDeclared(output)) { engine->DeclareOutput(output);
engine->DeclareOutput(output);
}
} }
engine->FreezeNetwork(); engine->FreezeNetwork();
} }
}; };
......
...@@ -24,8 +24,7 @@ limitations under the License. */ ...@@ -24,8 +24,7 @@ limitations under the License. */
#include "paddle/fluid/inference/tensorrt/convert/op_converter.h" #include "paddle/fluid/inference/tensorrt/convert/op_converter.h"
#include "paddle/fluid/inference/tensorrt/convert/ut_helper.h" #include "paddle/fluid/inference/tensorrt/convert/ut_helper.h"
USE_CUDA_ONLY_OP(tensorrt_engine); USE_NO_KERNEL_OP(tensorrt_engine);
namespace paddle { namespace paddle {
namespace operators { namespace operators {
......
/* Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License. */
#include "paddle/fluid/framework/data_layout_transform.h"
#include "paddle/fluid/framework/op_registry.h"
#include "paddle/fluid/memory/malloc.h"
#include "paddle/fluid/platform/mkldnn_reuse.h"
namespace paddle {
namespace operators {
using Tensor = framework::Tensor;
using framework::DataLayout;
template <typename T>
class TransposeMKLDNNOpKernel : public paddle::framework::OpKernel<T> {
public:
void Compute(const paddle::framework::ExecutionContext& ctx) const override {
PADDLE_ENFORCE(paddle::platform::is_cpu_place(ctx.GetPlace()),
"It must use CPUPlace.");
const bool is_test = ctx.Attr<bool>("is_test");
PADDLE_ENFORCE(
is_test == true,
"ConvTransposeMKLDNN works only for inference!. Set is_test = True");
auto& dev_ctx =
ctx.template device_context<paddle::platform::MKLDNNDeviceContext>();
const auto& mkldnn_engine = dev_ctx.GetEngine();
std::vector<int> axis = ctx.Attr<std::vector<int>>("axis");
int ndims = axis.size();
auto* input = ctx.Input<Tensor>("X");
auto* output = ctx.Output<Tensor>("Out");
const T* input_data = input->data<T>();
if (ndims == 1) {
output->ShareDataWith(*input);
return;
}
std::vector<int> nchw_axis(ndims, 0);
for (size_t i = 0; i < nchw_axis.size(); ++i) {
nchw_axis[i] = i;
}
std::vector<int> nchw_tz = paddle::framework::vectorize2int(input->dims());
std::string data_format = ctx.Attr<std::string>("data_format");
auto src_md =
input->format() != mkldnn::memory::format::nchw
? platform::MKLDNNMemDesc(nchw_tz, platform::MKLDNNGetDataType<T>(),
input->format())
: Axis2MemoryDesc(nchw_tz, nchw_axis);
this->TransposeKernel(ctx.GetPlace(), Axis2MemoryDesc(nchw_tz, axis),
src_md, output, input_data, nchw_tz, mkldnn_engine);
}
protected:
mkldnn::memory::desc Axis2MemoryDesc(std::vector<int>& nchw_tz,
std::vector<int>& axis) const {
mkldnn_memory_desc_t mem_fmt;
mem_fmt.primitive_kind = mkldnn_memory;
mem_fmt.ndims = axis.size();
for (unsigned int i = 0; i < nchw_tz.size(); ++i) {
mem_fmt.dims[i] = nchw_tz[i]; // logical dimensions (nchw format,
// regardless physical layout)
}
mem_fmt.data_type = mkldnn_f32;
mem_fmt.format = mkldnn_blocked;
unsigned int total_stride = 1;
for (int i = nchw_tz.size() - 1; i >= 0; --i) {
mem_fmt.layout_desc.blocking.padding_dims[i] =
nchw_tz[i]; // logical dimensions (nchw format, regardless physical
// layout)
mem_fmt.layout_desc.blocking.block_dims[i] = 1;
mem_fmt.layout_desc.blocking.offset_padding_to_data[i] = 0; // no offset
mem_fmt.layout_desc.blocking.strides[0][axis[i]] = total_stride;
mem_fmt.layout_desc.blocking.strides[1][axis[i]] = 1;
total_stride *= nchw_tz[axis[i]];
}
mem_fmt.layout_desc.blocking.offset_padding = 0; // no initial offset
return mem_fmt;
}
void TransposeKernel(platform::Place place, mkldnn::memory::desc md_o,
mkldnn::memory::desc md_i, Tensor* output,
const T* data_i, std::vector<int>& nchw_dims,
const mkldnn::engine& eng) const {
// Make Memory primitive descriptors
auto mpd_o = mkldnn::memory::primitive_desc(md_o, eng);
auto mpd_i = mkldnn::memory::primitive_desc(md_i, eng);
auto data_o = output->mutable_data<T>(
place, paddle::memory::Allocator::kDefault, mpd_o.get_size());
auto src = mkldnn::memory(mpd_i, (T*)(data_i));
auto dst = mkldnn::memory(mpd_o, data_o);
auto r = mkldnn::reorder(src, dst);
mkldnn::stream(mkldnn::stream::kind::eager).submit({r}).wait();
}
};
} // namespace operators
} // namespace paddle
namespace ops = paddle::operators;
REGISTER_OP_KERNEL(transpose2, MKLDNN, ::paddle::platform::CPUPlace,
ops::TransposeMKLDNNOpKernel<float>);
REGISTER_OP_KERNEL(transpose, MKLDNN, ::paddle::platform::CPUPlace,
ops::TransposeMKLDNNOpKernel<float>);
...@@ -16,6 +16,10 @@ limitations under the License. */ ...@@ -16,6 +16,10 @@ limitations under the License. */
#include <string> #include <string>
#include <vector> #include <vector>
#ifdef PADDLE_WITH_MKLDNN
#include "paddle/fluid/platform/mkldnn_helper.h"
#endif
namespace paddle { namespace paddle {
namespace operators { namespace operators {
...@@ -53,11 +57,32 @@ class TransposeOp : public framework::OperatorWithKernel { ...@@ -53,11 +57,32 @@ class TransposeOp : public framework::OperatorWithKernel {
} }
ctx->SetOutputDim("Out", out_dims); ctx->SetOutputDim("Out", out_dims);
} }
protected:
framework::OpKernelType GetExpectedKernelType(
const framework::ExecutionContext &ctx) const override {
framework::LibraryType library_{framework::LibraryType::kPlain};
std::string data_format = ctx.Attr<std::string>("data_format");
framework::DataLayout layout_ = framework::StringToDataLayout(data_format);
#ifdef PADDLE_WITH_MKLDNN
if (library_ == framework::LibraryType::kPlain &&
platform::CanMKLDNNBeUsed(ctx)) {
library_ = framework::LibraryType::kMKLDNN;
layout_ = framework::DataLayout::kMKLDNN;
}
#endif
return framework::OpKernelType(ctx.Input<Tensor>("X")->type(),
ctx.GetPlace(), layout_, library_);
}
}; };
class TransposeOpMaker : public framework::OpProtoAndCheckerMaker { class TransposeOpMaker : public framework::OpProtoAndCheckerMaker {
public: public:
void Make() override { void Make() override {
AddAttr<bool>("is_test",
"(bool, default false) Set to true for inference only, false "
"for training. Some layers may run faster when this is true.")
.SetDefault(false);
AddInput( AddInput(
"X", "X",
"(Tensor) The input tensor, tensors with rank up to 6 are supported."); "(Tensor) The input tensor, tensors with rank up to 6 are supported.");
...@@ -67,6 +92,16 @@ class TransposeOpMaker : public framework::OpProtoAndCheckerMaker { ...@@ -67,6 +92,16 @@ class TransposeOpMaker : public framework::OpProtoAndCheckerMaker {
"(vector<int>) A list of values, and the size of the list should be " "(vector<int>) A list of values, and the size of the list should be "
"the same with the input tensor rank. This operator permutes the input " "the same with the input tensor rank. This operator permutes the input "
"tensor's axes according to the values given."); "tensor's axes according to the values given.");
AddAttr<bool>("use_mkldnn",
"(bool, default false) Only used in mkldnn kernel")
.SetDefault(false);
AddAttr<std::string>(
"data_format",
"(string, default NCHW) Only used in "
"An optional string from: \"NHWC\", \"NCHW\". "
"Defaults to \"NHWC\". Specify the data format of the output data, "
"the input will be transformed automatically. ")
.SetDefault("AnyLayout");
AddComment(R"DOC( AddComment(R"DOC(
Transpose Operator. Transpose Operator.
...@@ -144,8 +179,18 @@ class Transpose2Op : public TransposeOp { ...@@ -144,8 +179,18 @@ class Transpose2Op : public TransposeOp {
protected: protected:
framework::OpKernelType GetExpectedKernelType( framework::OpKernelType GetExpectedKernelType(
const framework::ExecutionContext &ctx) const override { const framework::ExecutionContext &ctx) const override {
return framework::OpKernelType(ctx.Input<framework::LoDTensor>("X")->type(), framework::LibraryType library_{framework::LibraryType::kPlain};
ctx.device_context()); std::string data_format = ctx.Attr<std::string>("data_format");
framework::DataLayout layout_ = framework::StringToDataLayout(data_format);
#ifdef PADDLE_WITH_MKLDNN
if (library_ == framework::LibraryType::kPlain &&
platform::CanMKLDNNBeUsed(ctx)) {
library_ = framework::LibraryType::kMKLDNN;
layout_ = framework::DataLayout::kMKLDNN;
}
#endif
return framework::OpKernelType(ctx.Input<Tensor>("X")->type(),
ctx.GetPlace(), layout_, library_);
} }
}; };
......
...@@ -16,9 +16,7 @@ if (CUPTI_FOUND) ...@@ -16,9 +16,7 @@ if (CUPTI_FOUND)
list(APPEND CUDA_SRCS cupti.cc) list(APPEND CUDA_SRCS cupti.cc)
endif(CUPTI_FOUND) endif(CUPTI_FOUND)
nv_library(dynload_cuda SRCS ${CUDA_SRCS} DEPS dynamic_loader) nv_library(dynload_cuda SRCS ${CUDA_SRCS} DEPS dynamic_loader)
if (NOT WIN32)
cc_library(dynload_warpctc SRCS warpctc.cc DEPS dynamic_loader warpctc) cc_library(dynload_warpctc SRCS warpctc.cc DEPS dynamic_loader warpctc)
endif(NOT WIN32)
if (WITH_MKLML) if (WITH_MKLML)
cc_library(dynload_mklml SRCS mklml.cc DEPS dynamic_loader mklml) cc_library(dynload_mklml SRCS mklml.cc DEPS dynamic_loader mklml)
endif() endif()
......
...@@ -34,7 +34,7 @@ extern void EnforceCUDNNLoaded(const char* fn_name); ...@@ -34,7 +34,7 @@ extern void EnforceCUDNNLoaded(const char* fn_name);
#define DECLARE_DYNAMIC_LOAD_CUDNN_WRAP(__name) \ #define DECLARE_DYNAMIC_LOAD_CUDNN_WRAP(__name) \
struct DynLoad__##__name { \ struct DynLoad__##__name { \
template <typename... Args> \ template <typename... Args> \
auto operator()(Args... args) -> decltype(__name(args...)) { \ auto operator()(Args... args) -> DECLARE_TYPE(__name, args...) { \
using cudnn_func = decltype(&::__name); \ using cudnn_func = decltype(&::__name); \
std::call_once(cudnn_dso_flag, []() { \ std::call_once(cudnn_dso_flag, []() { \
cudnn_dso_handle = paddle::platform::dynload::GetCUDNNDsoHandle(); \ cudnn_dso_handle = paddle::platform::dynload::GetCUDNNDsoHandle(); \
......
...@@ -201,6 +201,8 @@ void* GetCurandDsoHandle() { ...@@ -201,6 +201,8 @@ void* GetCurandDsoHandle() {
void* GetWarpCTCDsoHandle() { void* GetWarpCTCDsoHandle() {
#if defined(__APPLE__) || defined(__OSX__) #if defined(__APPLE__) || defined(__OSX__)
return GetDsoHandleFromSearchPath(FLAGS_warpctc_dir, "libwarpctc.dylib"); return GetDsoHandleFromSearchPath(FLAGS_warpctc_dir, "libwarpctc.dylib");
#elif defined(_WIN32)
return GetDsoHandleFromSearchPath(FLAGS_warpctc_dir, "warpctc.dll");
#else #else
return GetDsoHandleFromSearchPath(FLAGS_warpctc_dir, "libwarpctc.so"); return GetDsoHandleFromSearchPath(FLAGS_warpctc_dir, "libwarpctc.so");
#endif #endif
......
...@@ -18,6 +18,12 @@ namespace paddle { ...@@ -18,6 +18,12 @@ namespace paddle {
namespace platform { namespace platform {
namespace dynload { namespace dynload {
#ifndef _WIN32
#define DECLARE_TYPE(__name, ...) decltype(__name(__VA_ARGS__))
#else
#define DECLARE_TYPE(__name, ...) decltype(auto)
#endif
void* GetCublasDsoHandle(); void* GetCublasDsoHandle();
void* GetCUDNNDsoHandle(); void* GetCUDNNDsoHandle();
void* GetCUPTIDsoHandle(); void* GetCUPTIDsoHandle();
......
...@@ -34,7 +34,7 @@ extern void* mklml_dso_handle; ...@@ -34,7 +34,7 @@ extern void* mklml_dso_handle;
#define DYNAMIC_LOAD_MKLML_WRAP(__name) \ #define DYNAMIC_LOAD_MKLML_WRAP(__name) \
struct DynLoad__##__name { \ struct DynLoad__##__name { \
template <typename... Args> \ template <typename... Args> \
auto operator()(Args... args) -> decltype(__name(args...)) { \ auto operator()(Args... args) -> DECLARE_TYPE(__name, args...) { \
using mklmlFunc = decltype(&::__name); \ using mklmlFunc = decltype(&::__name); \
std::call_once(mklml_dso_flag, []() { \ std::call_once(mklml_dso_flag, []() { \
mklml_dso_handle = paddle::platform::dynload::GetMKLMLDsoHandle(); \ mklml_dso_handle = paddle::platform::dynload::GetMKLMLDsoHandle(); \
...@@ -82,6 +82,8 @@ extern void* mklml_dso_handle; ...@@ -82,6 +82,8 @@ extern void* mklml_dso_handle;
__macro(vdSqr); \ __macro(vdSqr); \
__macro(vsPowx); \ __macro(vsPowx); \
__macro(vdPowx); \ __macro(vdPowx); \
__macro(vsInv); \
__macro(vdInv); \
__macro(MKL_Set_Num_Threads) __macro(MKL_Set_Num_Threads)
MKLML_ROUTINE_EACH(DECLARE_DYNAMIC_LOAD_MKLML_WRAP); MKLML_ROUTINE_EACH(DECLARE_DYNAMIC_LOAD_MKLML_WRAP);
......
...@@ -33,7 +33,7 @@ extern void* tensorrt_dso_handle; ...@@ -33,7 +33,7 @@ extern void* tensorrt_dso_handle;
#define DECLARE_DYNAMIC_LOAD_TENSORRT_WRAP(__name) \ #define DECLARE_DYNAMIC_LOAD_TENSORRT_WRAP(__name) \
struct DynLoad__##__name { \ struct DynLoad__##__name { \
template <typename... Args> \ template <typename... Args> \
auto operator()(Args... args) -> decltype(__name(args...)) { \ auto operator()(Args... args) -> DECLARE_TYPE(__name, args...) { \
using tensorrt_func = decltype(__name(args...)) (*)(Args...); \ using tensorrt_func = decltype(__name(args...)) (*)(Args...); \
std::call_once(tensorrt_dso_flag, []() { \ std::call_once(tensorrt_dso_flag, []() { \
tensorrt_dso_handle = \ tensorrt_dso_handle = \
......
...@@ -34,7 +34,7 @@ extern void* warpctc_dso_handle; ...@@ -34,7 +34,7 @@ extern void* warpctc_dso_handle;
#define DYNAMIC_LOAD_WARPCTC_WRAP(__name) \ #define DYNAMIC_LOAD_WARPCTC_WRAP(__name) \
struct DynLoad__##__name { \ struct DynLoad__##__name { \
template <typename... Args> \ template <typename... Args> \
auto operator()(Args... args) -> decltype(__name(args...)) { \ auto operator()(Args... args) -> DECLARE_TYPE(__name, args...) { \
using warpctcFunc = decltype(&::__name); \ using warpctcFunc = decltype(&::__name); \
std::call_once(warpctc_dso_flag, []() { \ std::call_once(warpctc_dso_flag, []() { \
warpctc_dso_handle = paddle::platform::dynload::GetWarpCTCDsoHandle(); \ warpctc_dso_handle = paddle::platform::dynload::GetWarpCTCDsoHandle(); \
......
...@@ -55,7 +55,6 @@ static void *dlsym(void *handle, const char *symbol_name) { ...@@ -55,7 +55,6 @@ static void *dlsym(void *handle, const char *symbol_name) {
static void *dlopen(const char *filename, int flag) { static void *dlopen(const char *filename, int flag) {
std::string file_name(filename); std::string file_name(filename);
file_name.replace(0, file_name.size() - 1, '/', '\\');
HMODULE hModule = LoadLibrary(file_name.c_str()); HMODULE hModule = LoadLibrary(file_name.c_str());
if (!hModule) { if (!hModule) {
throw std::runtime_error(file_name + " not found."); throw std::runtime_error(file_name + " not found.");
......
...@@ -960,6 +960,14 @@ All parameter, weight, gradient are variables in Paddle. ...@@ -960,6 +960,14 @@ All parameter, weight, gradient are variables in Paddle.
R"DOC(The type is BOOL, fuse_elewise_add_act_ops indicate whether R"DOC(The type is BOOL, fuse_elewise_add_act_ops indicate whether
to fuse elementwise_add_op and activation_op, to fuse elementwise_add_op and activation_op,
it may make the execution faster. Default False)DOC") it may make the execution faster. Default False)DOC")
.def_property(
"memory_optimize",
[](const BuildStrategy &self) { return self.memory_optimize_; },
[](BuildStrategy &self, bool b) { self.memory_optimize_ = b; })
.def_property(
"memory_early_delete",
[](const BuildStrategy &self) { return self.memory_early_delete_; },
[](BuildStrategy &self, bool b) { self.memory_early_delete_ = b; })
.def("_finalize_strategy_and_create_passes", .def("_finalize_strategy_and_create_passes",
[](BuildStrategy &self) -> std::shared_ptr<ir::PassBuilder> { [](BuildStrategy &self) -> std::shared_ptr<ir::PassBuilder> {
return self.CreatePassesFromStrategy(true); return self.CreatePassesFromStrategy(true);
......
...@@ -102,6 +102,13 @@ def __bootstrap__(): ...@@ -102,6 +102,13 @@ def __bootstrap__():
import sys import sys
import os import os
import platform import platform
if os.name == 'nt':
third_lib_path = os.path.abspath(os.path.dirname(
__file__)) + os.sep + '..' + os.sep + 'libs'
os.environ['path'] += ';' + third_lib_path
sys.path.append(third_lib_path)
from . import core from . import core
in_test = 'unittest' in sys.modules in_test = 'unittest' in sys.modules
...@@ -128,13 +135,12 @@ def __bootstrap__(): ...@@ -128,13 +135,12 @@ def __bootstrap__():
'free_idle_memory', 'paddle_num_threads', "dist_threadpool_size", 'free_idle_memory', 'paddle_num_threads', "dist_threadpool_size",
'eager_delete_tensor_gb', 'fast_eager_deletion_mode', 'eager_delete_tensor_gb', 'fast_eager_deletion_mode',
'allocator_strategy', 'reader_queue_speed_test_mode', 'allocator_strategy', 'reader_queue_speed_test_mode',
'print_sub_graph_dir', 'pe_profile_fname' 'print_sub_graph_dir', 'pe_profile_fname', 'warpctc_dir'
] ]
if 'Darwin' not in sysstr: if 'Darwin' not in sysstr:
read_env_flags.append('use_pinned_memory') read_env_flags.append('use_pinned_memory')
if os.name != 'nt': if os.name != 'nt':
read_env_flags.append('warpctc_dir')
read_env_flags.append('cpu_deterministic') read_env_flags.append('cpu_deterministic')
if core.is_compiled_with_dist(): if core.is_compiled_with_dist():
...@@ -150,7 +156,7 @@ def __bootstrap__(): ...@@ -150,7 +156,7 @@ def __bootstrap__():
read_env_flags += [ read_env_flags += [
'fraction_of_gpu_memory_to_use', 'cudnn_deterministic', 'fraction_of_gpu_memory_to_use', 'cudnn_deterministic',
'enable_cublas_tensor_op_math', 'conv_workspace_size_limit', 'enable_cublas_tensor_op_math', 'conv_workspace_size_limit',
'cudnn_exhaustive_search', 'selected_gpus' 'cudnn_exhaustive_search', 'memory_optimize_debug', 'selected_gpus'
] ]
core.init_gflags([sys.argv[0]] + core.init_gflags([sys.argv[0]] +
......
...@@ -249,69 +249,6 @@ def serialize_op_decs(op_desc): ...@@ -249,69 +249,6 @@ def serialize_op_decs(op_desc):
return proto.__str__() return proto.__str__()
def _callback_lookup_(op):
"""
Only used in _append_backward_ops_
Build and returns a callback function for certain op. For example
parallel_do: AllReduce
:param op:
:return: callback function
"""
if op.type == 'parallel_do' and op.attr('use_nccl'):
all_vars = op.block.vars
param_names = set(op.input('parameters'))
param_names = [
name for name in param_names
if all_vars[name].stop_gradient is False
]
param_grad_names = [n + "@GRAD" for n in param_names]
class ParallelDoCallBack(object):
def __init__(self, param_grad_names, parallel_scopes_name):
self.has_inserted_nccl_init = False
self.param_grad_names = param_grad_names
self.parallel_scopes_name = parallel_scopes_name
def __call__(self, block, context):
if not self.has_inserted_nccl_init:
op_desc = _create_op_desc_(
"ncclInit",
{"parallel_scopes": self.parallel_scopes_name},
{"Communicator": ['nccl_com__do_not_change_']}, {})
block.program.global_block().desc.append_op().copy_from(
op_desc)
self.has_inserted_nccl_init = True
current_op_desc = context["__current_op_desc__"]
for o_param in current_op_desc.output_names():
for o_argu in current_op_desc.output(o_param):
if o_argu in self.param_grad_names:
allreduce_out_name = o_argu + "__nccl_all_reduce__"
op_desc = _create_op_desc_(
"ncclReduce",
{
"X": [o_argu],
"Communicator":
['nccl_com__do_not_change_']
},
{"Out": [allreduce_out_name]},
{"reduction": "ncclSum",
"root": 0}, )
block.desc.append_op().copy_from(op_desc)
op_desc = _create_op_desc_(
"assign", {"X": [allreduce_out_name]},
{"Out": [o_argu]}, {})
block.desc.append_op().copy_from(op_desc)
return ParallelDoCallBack(param_grad_names,
op.output("parallel_scopes"))
else:
return None
def _append_backward_ops_(block, def _append_backward_ops_(block,
ops, ops,
target_block, target_block,
...@@ -349,17 +286,8 @@ def _append_backward_ops_(block, ...@@ -349,17 +286,8 @@ def _append_backward_ops_(block,
sub_block = program.block(op._block_attr_id("sub_block")) sub_block = program.block(op._block_attr_id("sub_block"))
grad_sub_block = program._create_block() grad_sub_block = program._create_block()
grad_sub_block._set_forward_block_idx(sub_block.idx) grad_sub_block._set_forward_block_idx(sub_block.idx)
cb = _callback_lookup_(op) _append_backward_ops_(sub_block, sub_block.ops, grad_sub_block,
if cb is not None: no_grad_dict, grad_to_var, callbacks)
if callbacks is None:
new_callbacks = [cb]
else:
new_callbacks = callbacks + [_callback_lookup_(op)]
_append_backward_ops_(sub_block, sub_block.ops, grad_sub_block,
no_grad_dict, grad_to_var, new_callbacks)
else:
_append_backward_ops_(sub_block, sub_block.ops, grad_sub_block,
no_grad_dict, grad_to_var, callbacks)
program._rollback() program._rollback()
grad_sub_block_list.append(grad_sub_block.desc) grad_sub_block_list.append(grad_sub_block.desc)
...@@ -424,9 +352,6 @@ def _append_backward_vars_(block, start_op_idx, grad_to_var, grad_info_map): ...@@ -424,9 +352,6 @@ def _append_backward_vars_(block, start_op_idx, grad_to_var, grad_info_map):
# infer_shape and infer_type # infer_shape and infer_type
op_desc.infer_var_type(block.desc) op_desc.infer_var_type(block.desc)
op_desc.infer_shape(block.desc) op_desc.infer_shape(block.desc)
# ncclInit dones't need to set data_type
if op_desc.type() == 'ncclInit':
continue
for arg in op_desc.output_arg_names(): for arg in op_desc.output_arg_names():
if arg in new_vars: if arg in new_vars:
_infer_var_data_type_(arg, block) _infer_var_data_type_(arg, block)
......
...@@ -16,6 +16,7 @@ from __future__ import print_function ...@@ -16,6 +16,7 @@ from __future__ import print_function
import collections import collections
import contextlib import contextlib
import os
import re import re
import six import six
import sys import sys
...@@ -27,11 +28,18 @@ from .proto import framework_pb2 ...@@ -27,11 +28,18 @@ from .proto import framework_pb2
try: try:
from . import core from . import core
except ImportError as e: except ImportError as e:
raise ImportError( if os.name == 'nt':
"""NOTE: You may need to run \"export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH\" raise ImportError(
if you encounters \"libmkldnn.so not found\" errors. If you have python """NOTE: You may need to run \"set PATH=c:\python27\lib:%PATH%\"
installed in other directory, replace \"/usr/local/lib\" with your own if you encounters \"mkldnn.dll not found\" errors. If you have python
directory. The original error is: \n""" + cpt.get_exception_message(e)) installed in other directory, replace \"c:\python27\lib" with your own
directory. The original error is: \n""" + cpt.get_exception_message(e))
else:
raise ImportError(
"""NOTE: You may need to run \"export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH\"
if you encounters \"libmkldnn.so not found\" errors. If you have python
installed in other directory, replace \"/usr/local/lib\" with your own
directory. The original error is: \n""" + cpt.get_exception_message(e))
except Exception as e: except Exception as e:
raise e raise e
from . import unique_name from . import unique_name
...@@ -563,8 +571,8 @@ class Operator(object): ...@@ -563,8 +571,8 @@ class Operator(object):
OP_WITHOUT_KERNEL_SET = { OP_WITHOUT_KERNEL_SET = {
'feed', 'fetch', 'save', 'load', 'recurrent', 'go', 'feed', 'fetch', 'save', 'load', 'recurrent', 'go',
'rnn_memory_helper_grad', 'conditional_block', 'while', 'send', 'recv', 'rnn_memory_helper_grad', 'conditional_block', 'while', 'send', 'recv',
'listen_and_serv', 'parallel_do', 'save_combine', 'load_combine', 'listen_and_serv', 'save_combine', 'load_combine', 'ncclInit', 'select',
'ncclInit', 'select', 'checkpoint_notify', 'gen_nccl_id' 'checkpoint_notify', 'gen_nccl_id'
} }
def __init__(self, def __init__(self,
......
...@@ -226,156 +226,6 @@ class BlockGuard(object): ...@@ -226,156 +226,6 @@ class BlockGuard(object):
return True return True
class ParallelDo(object):
"""
ParallelDo is used to represent multi-thread data parallel processing.
Its vanilla implementation can be shown as the following (:math:`|` means
single thread and :math:`||||` means multiple threads)
.. code-block:: text
In the forward pass
| Split input onto different devices
| Copy parameter onto different devices
|||| Compute forward pass in parallel
| Merge output from different devices
In the backward pass
| Split output@grad onto different devices
|||| Compute backward pass in parallel
| accumulate param@grad from different devices to the first device
| Merge input@grad from different devices
| Copy param@grad to the place of parallel_do_op
Examples:
.. code-block:: python
images = fluid.layers.data(name='pixel', shape=[1, 28, 28], dtype=DTYPE)
label = fluid.layers.data(name='label', shape=[1], dtype='int64')
# ParallelDo version & Single-thread version
if thread_num > 1:
places = fluid.layers.get_places(thread_num)
pd = fluid.layers.control_flow.ParallelDo(places)
with pd.do():
images = pd.read_input(images)
label = pd.read_input(label)
predict = cnn_model(images)
cost = fluid.layers.cross_entropy(input=predict, label=label)
avg_cost = fluid.layers.mean(x=cost)
pd.write_output(avg_cost)
avg_cost = pd()
avg_cost = fluid.layers.mean(avg_cost)
else:
predict = cnn_model(images)
cost = fluid.layers.cross_entropy(input=predict, label=label)
avg_cost = fluid.layers.mean(x=cost)
.. warning::
It will be soon deprecated, please use ParallelExecutor instead.
"""
def __init__(self, places, use_nccl=False, name=None):
warnings.warn(
"API ParallelDo is deprecated since 0.15.0. Please use ParallelExecutor instead.",
Warning)
self.helper = LayerHelper("parallel_do", name=name)
self.inputs = []
self.places = places
self.outputs = []
self.status = StaticRNN.BEFORE_RNN_BLOCK
self.use_nccl = use_nccl
def do(self):
return BlockGuardWithCompletion(self)
def parent_block(self):
prog = self.helper.main_program
parent_idx = prog.current_block().parent_idx
assert parent_idx >= 0
parent_block = prog.block(parent_idx)
return parent_block
def __call__(self, *args, **kwargs):
if self.status != StaticRNN.AFTER_RNN_BLOCK:
raise ValueError("RNN output can only be retrieved after rnn block")
if len(self.outputs) == 0:
raise ValueError("RNN has no output")
elif len(self.outputs) == 1:
return self.outputs[0]
else:
return self.outputs
def read_input(self, var):
self.inputs.append(var)
return var
def write_output(self, var):
self.outputs.append(var)
def get_parameters(self):
main_program = self.helper.main_program
current_block = main_program.current_block()
parent_block = self.parent_block()
local_inputs = set()
params = list()
for var in self.inputs:
local_inputs.add(var.name)
for op in current_block.ops:
for iname in op.input_names:
for in_var_name in op.input(iname):
if in_var_name not in local_inputs:
params.append(in_var_name)
for oname in op.output_names:
for out_var_name in op.output(oname):
local_inputs.add(out_var_name)
params = list(set(params))
return [parent_block.var(name) for name in params]
def _complete_op(self):
main_program = self.helper.main_program
current_block = main_program.current_block()
parent_block = self.parent_block()
step_scope = parent_block.create_var(
type=core.VarDesc.VarType.STEP_SCOPES)
self.outputs = [
parent_block.create_var(
name=o.name,
shape=o.shape,
dtype=o.dtype,
lod_level=o.lod_level,
persistable=o.persistable,
stop_gradient=o.stop_gradient) for o in self.outputs
]
inputs = [parent_block.var(i.name) for i in self.inputs]
outputs = [parent_block.var(o.name) for o in self.outputs]
parent_block.append_op(
type='parallel_do',
inputs={
'inputs': inputs,
'parameters': self.get_parameters(),
'places': self.places
},
outputs={'outputs': outputs,
'parallel_scopes': [step_scope]},
attrs={'sub_block': current_block,
'use_nccl': self.use_nccl})
class BlockGuardWithCompletion(BlockGuard): class BlockGuardWithCompletion(BlockGuard):
""" """
BlockGuardWithCompletion class. BlockGuardWithCompletion class.
...@@ -384,9 +234,8 @@ class BlockGuardWithCompletion(BlockGuard): ...@@ -384,9 +234,8 @@ class BlockGuardWithCompletion(BlockGuard):
""" """
def __init__(self, rnn): def __init__(self, rnn):
if not (isinstance(rnn, StaticRNN) or isinstance(rnn, ParallelDo)): if not isinstance(rnn, StaticRNN):
raise TypeError( raise TypeError("BlockGuardWithCompletion takes a StaticRNN")
"BlockGuardWithCompletion takes a StaticRNN or ParallelDo")
super(BlockGuardWithCompletion, self).__init__(rnn.helper.main_program) super(BlockGuardWithCompletion, self).__init__(rnn.helper.main_program)
self.rnn = rnn self.rnn = rnn
......
...@@ -177,6 +177,7 @@ __all__ = [ ...@@ -177,6 +177,7 @@ __all__ = [
'lstm', 'lstm',
'psroi_pool', 'psroi_pool',
'teacher_student_sigmoid_loss', 'teacher_student_sigmoid_loss',
'huber_loss',
] ]
kIgnoreIndex = -100 kIgnoreIndex = -100
...@@ -498,7 +499,7 @@ def lstm(input, ...@@ -498,7 +499,7 @@ def lstm(input,
If Device is GPU, This op will use cudnn LSTM implementation If Device is GPU, This op will use cudnn LSTM implementation
A four-gate Long Short-Term Memory network with no peephole connections. A four-gate Long Short-Term Memory network with no peephole connections.
In the forward pass the output ht and cell output ct for a given iteration can be computed from the recurrent input ht-1, In the forward pass the output ht and cell output ct for a given iteration can be computed from the recurrent input ht-1,
the cell input ct-1 and the previous layer input xt given matrices W, R and biases bW, bR from the following equations: the cell input ct-1 and the previous layer input xt given matrices W, R and biases bW, bR from the following equations:
$$ i_t = \\sigma(W_{ix}x_{t} + W_{ih}h_{t-1} + bx_i + bh_i) $$ $$ i_t = \\sigma(W_{ix}x_{t} + W_{ih}h_{t-1} + bx_i + bh_i) $$
...@@ -525,19 +526,19 @@ def lstm(input, ...@@ -525,19 +526,19 @@ def lstm(input,
- $\tilde{c_t}$ is also called candidate hidden state, - $\tilde{c_t}$ is also called candidate hidden state,
which is computed based on the current input and the previous hidden state. which is computed based on the current input and the previous hidden state.
Where sigmoid is the sigmoid operator: sigmoid(x) = 1 / (1 + e^-x), * represents a point-wise multiplication, Where sigmoid is the sigmoid operator: sigmoid(x) = 1 / (1 + e^-x), * represents a point-wise multiplication,
X represensts a matrix multiplication X represensts a matrix multiplication
Args: Args:
input (Variable): LSTM input tensor, shape MUST be ( seq_len x batch_size x input_size ) input (Variable): LSTM input tensor, shape MUST be ( seq_len x batch_size x input_size )
init_h(Variable): The initial hidden state of the LSTM init_h(Variable): The initial hidden state of the LSTM
This is a tensor with shape ( num_layers x batch_size x hidden_size) This is a tensor with shape ( num_layers x batch_size x hidden_size)
if is_bidirec = True, shape should be ( num_layers*2 x batch_size x hidden_size) if is_bidirec = True, shape should be ( num_layers*2 x batch_size x hidden_size)
init_c(Variable): The initial cell state of the LSTM. init_c(Variable): The initial cell state of the LSTM.
This is a tensor with shape ( num_layers x batch_size x hidden_size ) This is a tensor with shape ( num_layers x batch_size x hidden_size )
if is_bidirec = True, shape should be ( num_layers*2 x batch_size x hidden_size) if is_bidirec = True, shape should be ( num_layers*2 x batch_size x hidden_size)
max_len (int): max length of LSTM. the first dim of input tensor CAN NOT greater than max_len max_len (int): max length of LSTM. the first dim of input tensor CAN NOT greater than max_len
hidden_size (int): hidden size of the LSTM hidden_size (int): hidden size of the LSTM
num_layers (int): total layers number of the LSTM num_layers (int): total layers number of the LSTM
dropout_prob(float|0.0): dropout prob, dropout ONLY work between rnn layers, NOT between time steps dropout_prob(float|0.0): dropout prob, dropout ONLY work between rnn layers, NOT between time steps
...@@ -556,10 +557,10 @@ def lstm(input, ...@@ -556,10 +557,10 @@ def lstm(input,
if is_bidirec set to True, shape will be ( seq_len x batch_sze x hidden_size*2) if is_bidirec set to True, shape will be ( seq_len x batch_sze x hidden_size*2)
last_h(Tensor): the hidden state of the last step of LSTM last_h(Tensor): the hidden state of the last step of LSTM
shape is ( num_layers x batch_size x hidden_size ) shape is ( num_layers x batch_size x hidden_size )
if is_bidirec set to True, shape will be ( num_layers*2 x batch_size x hidden_size) if is_bidirec set to True, shape will be ( num_layers*2 x batch_size x hidden_size)
last_c(Tensor): the cell state of the last step of LSTM last_c(Tensor): the cell state of the last step of LSTM
shape is ( num_layers x batch_size x hidden_size ) shape is ( num_layers x batch_size x hidden_size )
if is_bidirec set to True, shape will be ( num_layers*2 x batch_size x hidden_size) if is_bidirec set to True, shape will be ( num_layers*2 x batch_size x hidden_size)
Examples: Examples:
...@@ -4659,7 +4660,7 @@ def ctc_greedy_decoder(input, blank, name=None): ...@@ -4659,7 +4660,7 @@ def ctc_greedy_decoder(input, blank, name=None):
[0.5, 0.1, 0.3, 0.1]] [0.5, 0.1, 0.3, 0.1]]
input.lod = [[4, 4]] input.lod = [[4, 4]]
Computation: Computation:
step1: Apply argmax to first input sequence which is input.data[0:4]. Then we get: step1: Apply argmax to first input sequence which is input.data[0:4]. Then we get:
...@@ -4692,7 +4693,7 @@ def ctc_greedy_decoder(input, blank, name=None): ...@@ -4692,7 +4693,7 @@ def ctc_greedy_decoder(input, blank, name=None):
Returns: Returns:
Variable: CTC greedy decode result which is a 2-D tensor with shape [Lp, 1]. Variable: CTC greedy decode result which is a 2-D tensor with shape [Lp, 1].
'Lp' is the sum if all output sequences' length. If all the sequences 'Lp' is the sum if all output sequences' length. If all the sequences
in result were empty, the result LoDTensor will be [-1] with in result were empty, the result LoDTensor will be [-1] with
LoD [[]] and dims [1, 1]. LoD [[]] and dims [1, 1].
Examples: Examples:
...@@ -5046,7 +5047,7 @@ def hsigmoid(input, ...@@ -5046,7 +5047,7 @@ def hsigmoid(input,
""" """
The hierarchical sigmoid operator is used to accelerate the training The hierarchical sigmoid operator is used to accelerate the training
process of language model. This operator organizes the classes into a process of language model. This operator organizes the classes into a
complete binary tree, or you can use is_custom to pass your own tree to complete binary tree, or you can use is_custom to pass your own tree to
implement hierarchical. Each leaf node represents a class(a word) and each implement hierarchical. Each leaf node represents a class(a word) and each
internal node acts as a binary classifier. For each word there's a unique internal node acts as a binary classifier. For each word there's a unique
path from root to it's leaf node, hsigmoid calculate the cost for each path from root to it's leaf node, hsigmoid calculate the cost for each
...@@ -5062,7 +5063,7 @@ def hsigmoid(input, ...@@ -5062,7 +5063,7 @@ def hsigmoid(input,
2. build a dict to store word_id -> word's leaf to root path, we call it path_table. 2. build a dict to store word_id -> word's leaf to root path, we call it path_table.
3. build a dict to store word_id -> code of word's leaf to root path, we call it path_code. Code 3. build a dict to store word_id -> code of word's leaf to root path, we call it path_code. Code
means label of each binary classification, using 1 indicate true, 0 indicate false. means label of each binary classification, using 1 indicate true, 0 indicate false.
4. now, each word should has its path and code along the path, you can pass a batch of path and code 4. now, each word should has its path and code along the path, you can pass a batch of path and code
related to the same batch of inputs. related to the same batch of inputs.
...@@ -5072,8 +5073,8 @@ def hsigmoid(input, ...@@ -5072,8 +5073,8 @@ def hsigmoid(input,
and :math:`D` is the feature size. and :math:`D` is the feature size.
label (Variable): The tensor variable contains labels of training data. label (Variable): The tensor variable contains labels of training data.
It's a tensor with shape is :math:`[N \\times 1]`. It's a tensor with shape is :math:`[N \\times 1]`.
num_classes: (int), The number of classes, must not be less than 2. with default tree this has to be set, num_classes: (int), The number of classes, must not be less than 2. with default tree this has to be set,
it should never be None under is_custom=False, but while is_custom is true, it should be non leaf num it should never be None under is_custom=False, but while is_custom is true, it should be non leaf num
which indicates the num of classes using by binary classify. which indicates the num of classes using by binary classify.
param_attr (ParamAttr|None): The parameter attribute for learnable parameters/weights param_attr (ParamAttr|None): The parameter attribute for learnable parameters/weights
of hsigmoid. If it is set to None or one attribute of ParamAttr, hsigmoid of hsigmoid. If it is set to None or one attribute of ParamAttr, hsigmoid
...@@ -5086,15 +5087,15 @@ def hsigmoid(input, ...@@ -5086,15 +5087,15 @@ def hsigmoid(input,
is not set, the bias is initialized zero. Default: None. is not set, the bias is initialized zero. Default: None.
name (str|None): A name for this layer(optional). If set None, the layer name (str|None): A name for this layer(optional). If set None, the layer
will be named automatically. Default: None. will be named automatically. Default: None.
path_table: (Variable|None) this variable can store each batch of samples' path to root, path_table: (Variable|None) this variable can store each batch of samples' path to root,
it should be in leaf -> root order it should be in leaf -> root order
path_table should have the same shape with path_code, and for each sample i path_table[i] indicates a np.array like path_table should have the same shape with path_code, and for each sample i path_table[i] indicates a np.array like
structure and each element in this array is indexes in parent nodes' Weight Matrix. structure and each element in this array is indexes in parent nodes' Weight Matrix.
path_code: (Variable|None) this variable can store each batch of samples' code, path_code: (Variable|None) this variable can store each batch of samples' code,
each code consist with every code of parent nodes. it should be in leaf -> root order each code consist with every code of parent nodes. it should be in leaf -> root order
is_custom: (bool|False)using user defined binary tree instead of default complete binary tree, if costum is is_custom: (bool|False)using user defined binary tree instead of default complete binary tree, if costum is
set you need to set path_table/path_code/num_classes, otherwise num_classes should be set set you need to set path_table/path_code/num_classes, otherwise num_classes should be set
is_sparse: (bool|False)using sparse update instead of dense update, if set, the gradient is_sparse: (bool|False)using sparse update instead of dense update, if set, the gradient
of W and input will be sparse. of W and input will be sparse.
Returns: Returns:
...@@ -9419,3 +9420,51 @@ def psroi_pool(input, ...@@ -9419,3 +9420,51 @@ def psroi_pool(input,
'pooled_width': pooled_width 'pooled_width': pooled_width
}) })
return out return out
def huber_loss(input, label, delta):
"""
Huber loss is a loss function used in robust.
Huber loss can evaluate the fitness of input to label.
Different from MSE loss, Huber loss is more robust for outliers.
When the difference between input and label is large than delta
.. math::
huber\_loss = delta * (label - input) - 0.5 * delta * delta
When the difference between input and label is less than delta
.. math::
huber\_loss = 0.5 * (label - input) * (label - input)
Args:
input (Variable): This input is a probability computed by the previous operator.
The first dimension is batch size, and the last dimension is 1.
label (Variable): The groud truth whose first dimension is batch size
and last dimension is 1.
delta (float): The parameter of huber loss, which controls
the range of outliers
Returns:
huber\_loss (Variable): The huber loss with shape [batch_size, 1].
Examples:
.. code-block:: python
predictions = fluid.layers.softmax(x)
loss = fluid.layers.huber_loss(input=predictions, label=label, 1.0)
"""
helper = LayerHelper('huber_loss', **locals())
residual = helper.create_variable_for_type_inference(
dtype=helper.input_dtype())
out = helper.create_variable_for_type_inference(dtype=helper.input_dtype())
helper.append_op(
type='huber_loss',
inputs={'X': input,
'Y': label},
outputs={'Out': out,
'Residual': residual},
attrs={'delta': delta})
return out
...@@ -641,9 +641,14 @@ class AdamOptimizer(Optimizer): ...@@ -641,9 +641,14 @@ class AdamOptimizer(Optimizer):
beta1 (float): The exponential decay rate for the 1st moment estimates. beta1 (float): The exponential decay rate for the 1st moment estimates.
beta2 (float): The exponential decay rate for the 2nd moment estimates. beta2 (float): The exponential decay rate for the 2nd moment estimates.
epsilon (float): a small float value for numerical stability. epsilon (float): a small float value for numerical stability.
regularization: A Regularizer, such as regularization: A Regularizer, such as fluid.regularizer.L2DecayRegularizer.
fluid.regularizer.L2DecayRegularizer.
name: A optional name prefix. name: A optional name prefix.
lazy_mode(bool: false): The official Adam algorithm has two moving-average accumulators
the accumulators are updated at every step. Every element of the two moving-average is updated
in both dense mode and sparse mode. If the size of parameter is very large, then the update
may be very slow. The lazy mode only update the element that has gradient is the current
mini-batch, so it will be much more faster. But this mode has different semantics with the
original Adam algorithm and may lead to different result.
Examples: Examples:
.. code-block:: python .. code-block:: python
...@@ -663,7 +668,8 @@ class AdamOptimizer(Optimizer): ...@@ -663,7 +668,8 @@ class AdamOptimizer(Optimizer):
beta2=0.999, beta2=0.999,
epsilon=1e-8, epsilon=1e-8,
regularization=None, regularization=None,
name=None): name=None,
lazy_mode=False):
assert learning_rate is not None assert learning_rate is not None
assert beta1 is not None assert beta1 is not None
assert beta2 is not None assert beta2 is not None
...@@ -676,6 +682,7 @@ class AdamOptimizer(Optimizer): ...@@ -676,6 +682,7 @@ class AdamOptimizer(Optimizer):
self._beta1 = beta1 self._beta1 = beta1
self._beta2 = beta2 self._beta2 = beta2
self._epsilon = epsilon self._epsilon = epsilon
self._lazy_mode = lazy_mode
def _create_accumulators(self, block, parameters): def _create_accumulators(self, block, parameters):
assert isinstance(block, framework.Block) assert isinstance(block, framework.Block)
...@@ -729,7 +736,8 @@ class AdamOptimizer(Optimizer): ...@@ -729,7 +736,8 @@ class AdamOptimizer(Optimizer):
attrs={ attrs={
"beta1": self._beta1, "beta1": self._beta1,
"beta2": self._beta2, "beta2": self._beta2,
"epsilon": self._epsilon "epsilon": self._epsilon,
"lazy_mode": self._lazy_mode
}) })
return adam_op return adam_op
......
...@@ -15,7 +15,6 @@ ...@@ -15,7 +15,6 @@
from __future__ import print_function from __future__ import print_function
from paddle.fluid.layers.device import get_places from paddle.fluid.layers.device import get_places
from paddle.fluid.layers.control_flow import ParallelDo
import unittest import unittest
import paddle.fluid as fluid import paddle.fluid as fluid
import paddle import paddle
...@@ -147,22 +146,7 @@ def train(word_dict, ...@@ -147,22 +146,7 @@ def train(word_dict,
cost, acc_out, prediction = net_method( cost, acc_out, prediction = net_method(
data, label, input_dim=dict_dim, class_dim=class_dim) data, label, input_dim=dict_dim, class_dim=class_dim)
else: else:
places = get_places() raise NotImplementedError()
pd = ParallelDo(places)
with pd.do():
cost, acc, _ = net_method(
pd.read_input(data),
pd.read_input(label),
input_dim=dict_dim,
class_dim=class_dim)
pd.write_output(cost)
pd.write_output(acc)
cost, acc = pd()
cost = fluid.layers.mean(cost)
acc_out = fluid.layers.mean(acc)
prediction = None
assert save_dirname is None
adagrad = fluid.optimizer.Adagrad(learning_rate=0.002) adagrad = fluid.optimizer.Adagrad(learning_rate=0.002)
adagrad.minimize(cost) adagrad.minimize(cost)
......
...@@ -25,7 +25,6 @@ import numpy ...@@ -25,7 +25,6 @@ import numpy
import paddle import paddle
import paddle.fluid as fluid import paddle.fluid as fluid
from paddle.fluid.layers.device import get_places from paddle.fluid.layers.device import get_places
from paddle.fluid.layers.control_flow import ParallelDo
BATCH_SIZE = 64 BATCH_SIZE = 64
...@@ -82,19 +81,7 @@ def train(nn_type, ...@@ -82,19 +81,7 @@ def train(nn_type,
net_conf = conv_net net_conf = conv_net
if parallel: if parallel:
places = get_places() raise NotImplementedError()
pd = ParallelDo(places)
with pd.do():
img_ = pd.read_input(img)
label_ = pd.read_input(label)
prediction, avg_loss, acc = net_conf(img_, label_)
for o in [avg_loss, acc]:
pd.write_output(o)
avg_loss, acc = pd()
# get mean loss and acc through every devices.
avg_loss = fluid.layers.mean(avg_loss)
acc = fluid.layers.mean(acc)
else: else:
prediction, avg_loss, acc = net_conf(img, label) prediction, avg_loss, acc = net_conf(img, label)
...@@ -273,7 +260,7 @@ def inject_all_tests(): ...@@ -273,7 +260,7 @@ def inject_all_tests():
for use_cuda in (False, True): for use_cuda in (False, True):
if use_cuda and not core.is_compiled_with_cuda(): if use_cuda and not core.is_compiled_with_cuda():
continue continue
for parallel in (False, True): for parallel in (False, ):
for nn_type in ('mlp', 'conv'): for nn_type in ('mlp', 'conv'):
inject_test_method(use_cuda, parallel, nn_type, True) inject_test_method(use_cuda, parallel, nn_type, True)
......
...@@ -17,7 +17,6 @@ from __future__ import print_function ...@@ -17,7 +17,6 @@ from __future__ import print_function
import paddle import paddle
import paddle.fluid as fluid import paddle.fluid as fluid
from paddle.fluid.layers.device import get_places from paddle.fluid.layers.device import get_places
from paddle.fluid.layers.control_flow import ParallelDo
import unittest import unittest
import os import os
import numpy as np import numpy as np
...@@ -84,18 +83,7 @@ def train(use_cuda, is_sparse, is_parallel, save_dirname, is_local=True): ...@@ -84,18 +83,7 @@ def train(use_cuda, is_sparse, is_parallel, save_dirname, is_local=True):
avg_cost, predict_word = __network__( avg_cost, predict_word = __network__(
[first_word, second_word, third_word, forth_word, next_word]) [first_word, second_word, third_word, forth_word, next_word])
else: else:
places = get_places() raise NotImplementedError()
pd = ParallelDo(places)
with pd.do():
avg_cost, predict_word = __network__(
list(
map(pd.read_input, [
first_word, second_word, third_word, forth_word,
next_word
])))
pd.write_output(avg_cost)
avg_cost = fluid.layers.mean(pd())
sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001) sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
sgd_optimizer.minimize(avg_cost) sgd_optimizer.minimize(avg_cost)
...@@ -262,7 +250,7 @@ def inject_test_method(use_cuda, is_sparse, is_parallel): ...@@ -262,7 +250,7 @@ def inject_test_method(use_cuda, is_sparse, is_parallel):
for use_cuda in (False, True): for use_cuda in (False, True):
for is_sparse in (False, True): for is_sparse in (False, True):
for is_parallel in (False, True): for is_parallel in (False, ):
inject_test_method(use_cuda, is_sparse, is_parallel) inject_test_method(use_cuda, is_sparse, is_parallel)
if __name__ == '__main__': if __name__ == '__main__':
......
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import print_function
import math
import sys
import paddle
import paddle.fluid as fluid
from paddle.fluid.layers.device import get_places
from paddle.fluid.layers.control_flow import ParallelDo
# need to fix random seed and training data to compare the loss
# value accurately calculated by the default and the memory optimization
# version.
fluid.default_startup_program().random_seed = 111
x = fluid.layers.data(name='x', shape=[13], dtype='float32')
y = fluid.layers.data(name='y', shape=[1], dtype='float32')
device_type = 'CPU'
use_nccl = False
place = fluid.CPUPlace()
if fluid.core.is_compiled_with_cuda():
device_type = 'CUDA'
use_nccl = False
place = fluid.CUDAPlace(0)
places = get_places(device_count=0, device_type=device_type)
pd = ParallelDo(places, use_nccl=use_nccl)
with pd.do():
x_ = pd.read_input(x)
y_ = pd.read_input(y)
y_predict = fluid.layers.fc(input=x_, size=1, act=None)
cost = fluid.layers.square_error_cost(input=y_predict, label=y_)
avg_cost = fluid.layers.mean(x=cost)
pd.write_output(avg_cost)
cost = pd()
avg_cost = fluid.layers.mean(x=cost)
sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.01)
sgd_optimizer.minimize(avg_cost)
fluid.memory_optimize(fluid.default_main_program(), print_log=True)
# fluid.release_memory(fluid.default_main_program())
BATCH_SIZE = 200
# fix the order of training data
train_reader = paddle.batch(
paddle.dataset.uci_housing.train(), batch_size=BATCH_SIZE, drop_last=False)
# train_reader = paddle.batch(
# paddle.reader.shuffle(
# paddle.dataset.uci_housing.train(), buf_size=500),
# batch_size=BATCH_SIZE)
feeder = fluid.DataFeeder(place=place, feed_list=[x, y])
exe = fluid.Executor(place)
exe.run(fluid.default_startup_program())
PASS_NUM = 100
for pass_id in range(PASS_NUM):
for data in train_reader():
avg_loss_value, = exe.run(fluid.default_main_program(),
feed=feeder.feed(data),
fetch_list=[avg_cost])
if avg_loss_value[0] < 10.0:
exit(0) # if avg cost less than 10.0, we think our code is good.
print(avg_loss_value[0])
if math.isnan(float(avg_loss_value)):
sys.exit("got NaN loss, training failed.")
exit(1)
...@@ -39,6 +39,7 @@ class TestParallelExecutorBase(unittest.TestCase): ...@@ -39,6 +39,7 @@ class TestParallelExecutorBase(unittest.TestCase):
seed=None, seed=None,
use_parallel_executor=True, use_parallel_executor=True,
use_reduce=False, use_reduce=False,
use_ir_memory_optimize=False,
fuse_elewise_add_act_ops=False, fuse_elewise_add_act_ops=False,
optimizer=fluid.optimizer.Adam, optimizer=fluid.optimizer.Adam,
use_fast_executor=False, use_fast_executor=False,
...@@ -82,6 +83,7 @@ class TestParallelExecutorBase(unittest.TestCase): ...@@ -82,6 +83,7 @@ class TestParallelExecutorBase(unittest.TestCase):
build_strategy.reduce_strategy = fluid.BuildStrategy.ReduceStrategy.Reduce \ build_strategy.reduce_strategy = fluid.BuildStrategy.ReduceStrategy.Reduce \
if use_reduce else fluid.BuildStrategy.ReduceStrategy.AllReduce if use_reduce else fluid.BuildStrategy.ReduceStrategy.AllReduce
build_strategy.fuse_elewise_add_act_ops = fuse_elewise_add_act_ops build_strategy.fuse_elewise_add_act_ops = fuse_elewise_add_act_ops
build_strategy.memory_optimize = use_ir_memory_optimize
build_strategy.enable_sequential_execution = enable_sequential_execution build_strategy.enable_sequential_execution = enable_sequential_execution
if use_cuda and core.is_compiled_with_cuda(): if use_cuda and core.is_compiled_with_cuda():
build_strategy.remove_unnecessary_lock = True build_strategy.remove_unnecessary_lock = True
......
...@@ -194,7 +194,8 @@ def adam_step(inputs, attributes): ...@@ -194,7 +194,8 @@ def adam_step(inputs, attributes):
return param_out, moment1_out, moment2_out return param_out, moment1_out, moment2_out
def adam_step_sparse(inputs, attributes, height, rows, row_numel, np_grad): def adam_step_sparse(inputs, attributes, height, rows, row_numel, np_grad,
lazy_mode):
''' '''
Simulate one step of the adam optimizer Simulate one step of the adam optimizer
:param inputs: dict of inputs :param inputs: dict of inputs
...@@ -218,19 +219,30 @@ def adam_step_sparse(inputs, attributes, height, rows, row_numel, np_grad): ...@@ -218,19 +219,30 @@ def adam_step_sparse(inputs, attributes, height, rows, row_numel, np_grad):
moment2_out = np.zeros(shape=[height, row_numel]) moment2_out = np.zeros(shape=[height, row_numel])
param_out = np.zeros(shape=[height, row_numel]) param_out = np.zeros(shape=[height, row_numel])
for idx, row_id in enumerate(rows): def update_row(row_id, update_value):
moment1_out[row_id] = beta1 * moment1[row_id] + (1 - beta1 moment1_out[row_id] = beta1 * moment1[row_id] + (1 - beta1
) * np_grad[idx] ) * update_value
moment2_out[row_id] = beta2 * moment2[row_id] + ( moment2_out[row_id] = beta2 * moment2[row_id] + (
1 - beta2) * np.square(np_grad[idx]) 1 - beta2) * np.square(update_value)
lr_t = lr * np.sqrt(1 - beta2_pow) / (1 - beta1_pow) lr_t = lr * np.sqrt(1 - beta2_pow) / (1 - beta1_pow)
param_out[row_id] = param[row_id] - lr_t * (moment1_out[row_id] / ( param_out[row_id] = param[row_id] - lr_t * (moment1_out[row_id] / (
np.sqrt(moment2_out[row_id]) + epsilon)) np.sqrt(moment2_out[row_id]) + epsilon))
if lazy_mode:
for idx, row_id in enumerate(rows):
update_row(row_id, np_grad[idx])
else:
for row_id in range(param_out.shape[0]):
update_value = np.zeros(np_grad[0].shape).astype("float32")
if row_id in rows:
update_value = np_grad[rows.index(row_id)]
update_row(row_id, update_value)
return param_out, moment1_out, moment2_out return param_out, moment1_out, moment2_out
class TestSparseAdamOp(unittest.TestCase): class TestSparseAdamOp(unittest.TestCase):
def setup(self, scope, place): def setup(self, scope, place, lazy_mode):
beta1 = 0.78 beta1 = 0.78
beta2 = 0.836 beta2 = 0.836
epsilon = 1e-4 epsilon = 1e-4
...@@ -248,6 +260,7 @@ class TestSparseAdamOp(unittest.TestCase): ...@@ -248,6 +260,7 @@ class TestSparseAdamOp(unittest.TestCase):
'Beta2Pow': np.array([beta2**10]).astype("float32"), 'Beta2Pow': np.array([beta2**10]).astype("float32"),
"LearningRate": np.full((1), 2.0).astype("float32") "LearningRate": np.full((1), 2.0).astype("float32")
} }
self.init_output = np.full((height, row_numel), 0.0).astype("float32")
self.attrs = {'epsilon': epsilon, 'beta1': beta1, 'beta2': beta2} self.attrs = {'epsilon': epsilon, 'beta1': beta1, 'beta2': beta2}
grad_selected_rows = scope.var('Grad').get_selected_rows() grad_selected_rows = scope.var('Grad').get_selected_rows()
...@@ -262,19 +275,21 @@ class TestSparseAdamOp(unittest.TestCase): ...@@ -262,19 +275,21 @@ class TestSparseAdamOp(unittest.TestCase):
self.sparse_inputs = ["Grad"] self.sparse_inputs = ["Grad"]
param_out, mom1, mom2 = adam_step_sparse( param_out, mom1, mom2 = adam_step_sparse(self.dense_inputs, self.attrs,
self.dense_inputs, self.attrs, height, rows, row_numel, np_array) height, rows, row_numel,
np_array, lazy_mode)
self.outputs = { self.outputs = {
"ParamOut": param_out, "ParamOut": param_out,
"Moment1Out": mom1, "Moment1Out": mom1,
"Moment2Out": mom2 "Moment2Out": mom2
} }
def check_with_place(self, place): def check_with_place(self, place, lazy_mode):
scope = core.Scope() scope = core.Scope()
self.setup(scope, place) self.setup(scope, place, lazy_mode)
op_args = dict() op_args = dict()
op_args['lazy_mode'] = lazy_mode
for key, np_array in self.dense_inputs.items(): for key, np_array in self.dense_inputs.items():
var = scope.var(key).get_tensor() var = scope.var(key).get_tensor()
var.set(np_array, place) var.set(np_array, place)
...@@ -283,7 +298,7 @@ class TestSparseAdamOp(unittest.TestCase): ...@@ -283,7 +298,7 @@ class TestSparseAdamOp(unittest.TestCase):
op_args[s] = s op_args[s] = s
for s in self.outputs: for s in self.outputs:
var = scope.var(s).get_tensor() var = scope.var(s).get_tensor()
var.set(self.outputs[s], place) var.set(self.init_output, place)
op_args[s] = s op_args[s] = s
for k in self.attrs: for k in self.attrs:
op_args[k] = self.attrs[k] op_args[k] = self.attrs[k]
...@@ -297,20 +312,17 @@ class TestSparseAdamOp(unittest.TestCase): ...@@ -297,20 +312,17 @@ class TestSparseAdamOp(unittest.TestCase):
actual = np.array(out_var) actual = np.array(out_var)
actual = actual.reshape([actual.size]) actual = actual.reshape([actual.size])
np_array = np_array.reshape([np_array.size]) np_array = np_array.reshape([np_array.size])
for idx, row_id in enumerate(self.rows):
j = 0 for i in range(np_array.size):
while j < self.row_numel: self.assertLess((actual[i] - np_array[i]), 0.00001)
pos = row_id * self.row_numel + j
self.assertLess((actual[pos] - np_array[pos]) / actual[pos], def test_sparse_adam(self):
0.00001)
j += 1
def test_sparse_sgd(self):
places = [core.CPUPlace()] places = [core.CPUPlace()]
if core.is_compiled_with_cuda(): if core.is_compiled_with_cuda():
places.append(core.CUDAPlace(0)) places.append(core.CUDAPlace(0))
for place in places: for place in places:
self.check_with_place(place) for lazy_mode in (True, False):
self.check_with_place(place, lazy_mode)
if __name__ == "__main__": if __name__ == "__main__":
......
...@@ -39,6 +39,7 @@ def train(network, use_cuda, use_parallel_executor, batch_size=32, pass_num=2): ...@@ -39,6 +39,7 @@ def train(network, use_cuda, use_parallel_executor, batch_size=32, pass_num=2):
label = fluid.layers.data(name="label", shape=[1], dtype="int64") label = fluid.layers.data(name="label", shape=[1], dtype="int64")
cost = network(data, label, len(word_dict)) cost = network(data, label, len(word_dict))
cost.persistable = True
optimizer = fluid.optimizer.Adagrad(learning_rate=0.2) optimizer = fluid.optimizer.Adagrad(learning_rate=0.2)
optimizer.minimize(cost) optimizer.minimize(cost)
......
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from parallel_executor_test_base import TestParallelExecutorBase
import paddle.fluid as fluid
import paddle.fluid.core as core
import numpy as np
import paddle
import paddle.dataset.mnist as mnist
import unittest
import os
MNIST_RECORDIO_FILE = "./mnist_test_pe.recordio"
def _feed_data_helper(use_feed):
if use_feed:
img = fluid.layers.data(name='image', shape=[784], dtype='float32')
label = fluid.layers.data(name='label', shape=[1], dtype='int64')
else:
reader = fluid.layers.open_files(
filenames=[MNIST_RECORDIO_FILE],
shapes=[[-1, 784], [-1, 1]],
lod_levels=[0, 0],
dtypes=['float32', 'int64'])
reader = fluid.layers.io.double_buffer(reader)
img, label = fluid.layers.read_file(reader)
return img, label
def simple_fc_net(use_feed):
x, y = _feed_data_helper(use_feed)
hidden_layer = 4
for _ in range(hidden_layer):
x = fluid.layers.fc(input=x, size=20, act='relu')
y_predict = fluid.layers.fc(input=x, size=10, act='softmax')
cost = fluid.layers.cross_entropy(input=y_predict, label=y)
avg_cost = fluid.layers.mean(cost)
return avg_cost
def fc_with_inplace_net(use_feed):
x, y = _feed_data_helper(use_feed)
fc = fluid.layers.fc(input=x, size=20, act='relu')
fc = fluid.layers.fc(input=fc, size=10, act='relu')
reshape = fluid.layers.reshape(x=fc, shape=[-1, 2, 5])
reshape = fluid.layers.reshape(x=reshape, shape=[-1, 5, 2])
y_predict = fluid.layers.fc(input=reshape, size=10, act='softmax')
cost = fluid.layers.cross_entropy(input=y_predict, label=y)
avg_cost = fluid.layers.mean(cost)
return avg_cost
class TestMNIST(TestParallelExecutorBase):
@classmethod
def setUpClass(cls):
os.environ['CPU_NUM'] = str(4)
# Convert mnist to recordio file
with fluid.program_guard(fluid.Program(), fluid.Program()):
reader = paddle.batch(mnist.train(), batch_size=4)
feeder = fluid.DataFeeder(
feed_list=[ # order is image and label
fluid.layers.data(
name='image', shape=[784]),
fluid.layers.data(
name='label', shape=[1], dtype='int64'),
],
place=fluid.CPUPlace())
fluid.recordio_writer.convert_reader_to_recordio_file(
MNIST_RECORDIO_FILE, reader, feeder)
def _dummy_data(self):
np.random.seed(5)
img = np.random.random(size=[32, 784]).astype(np.float32)
label = np.ones(shape=[32, 1], dtype='int64')
return img, label
def _compare_ir_and_python_memory_optimize(self, model, use_cuda):
if use_cuda and not core.is_compiled_with_cuda():
return
img, label = self._dummy_data()
first_loss0, last_loss0 = self.check_network_convergence(
model,
feed_dict={"image": img,
"label": label},
use_cuda=use_cuda,
memory_opt=False,
use_ir_memory_optimize=False)
first_loss1, last_loss1 = self.check_network_convergence(
model,
feed_dict={"image": img,
"label": label},
use_cuda=use_cuda,
memory_opt=False,
use_ir_memory_optimize=True)
for loss in zip(first_loss0, first_loss1):
self.assertAlmostEqual(loss[0], loss[1], delta=1e-6)
for loss in zip(last_loss0, last_loss1):
self.assertAlmostEqual(loss[0], loss[1], delta=1e-6)
def test_simple_fc_net(self):
self._compare_ir_and_python_memory_optimize(simple_fc_net, False)
self._compare_ir_and_python_memory_optimize(simple_fc_net, True)
def test_fc_with_reshape_net(self):
self._compare_ir_and_python_memory_optimize(fc_with_inplace_net, False)
self._compare_ir_and_python_memory_optimize(fc_with_inplace_net, True)
if __name__ == '__main__':
unittest.main()
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import print_function
import unittest
import paddle.fluid as fluid
from paddle.fluid.layers.device import get_places
from paddle.fluid.layers.control_flow import ParallelDo
import paddle.fluid.profiler as profiler
import numpy
import six
class BaseParallelForTest(unittest.TestCase):
def run_test(self, callback, feed, fetch):
"""
Run the unittest for parallel.for
Args:
callback(callable): A callable function returns a generator. There
are two yields in the generator function. The first yield
returns the data layers, and the second yield returns the loss.
The modified data variables will be sent back during the first
yield.
feed(dict): The executor feeding dictionary.
fetch(list|basestr): The fetch name lists.
Returns:
None
Raises:
AssertionError when the computation of cpu, parallel.for in cpu,
gpu, parallel.for in gpu are different.
"""
cpu = fluid.CPUPlace()
result_cpu = self._run_test_impl_(
callback=callback,
feed=feed,
fetch=fetch,
place=cpu,
use_parallel=False)
result_cpu_parallel = self._run_test_impl_(
callback=callback,
feed=feed,
fetch=fetch,
place=cpu,
use_parallel=True)
if fluid.core.is_compiled_with_cuda():
gpu = fluid.CUDAPlace(0)
result_gpu = self._run_test_impl_(
callback=callback,
feed=feed,
fetch=fetch,
place=gpu,
use_parallel=False,
use_gpu=True)
result_gpu_parallel = self._run_test_impl_(
callback=callback,
feed=feed,
fetch=fetch,
place=gpu,
use_parallel=True,
use_gpu=True)
result_gpu_nccl = self._run_test_impl_(
callback=callback,
feed=feed,
fetch=fetch,
place=gpu,
use_parallel=True,
use_nccl=True,
use_gpu=True)
self._assert_same_(fetch, result_cpu, result_cpu_parallel,
result_gpu, result_gpu_parallel, result_gpu_nccl)
else:
self._assert_same_(fetch, result_cpu, result_cpu_parallel)
def _run_test_impl_(self,
callback,
feed,
fetch,
place,
use_parallel=False,
use_nccl=False,
use_gpu=False):
"""
Run a single test, returns the fetch values
Args:
place(Place): the computation place.
use_parallel(bool): Whether use parallel.for or not.
Returns:
Fetched numpy arrays.
"""
if isinstance(fetch, six.string_types):
fetch = [fetch]
main = fluid.Program()
startup = fluid.Program()
# Fix seed
main.random_seed = 10
startup.random_seed = 10
with fluid.program_guard(main, startup):
generator = callback()
# Automatically insert parallel do if use_parallel = True
if use_parallel:
thread_num = fluid.core.get_cuda_device_count(
) if use_gpu else 8
places = get_places(thread_num)
pd = ParallelDo(places, use_nccl=use_nccl)
data = next(generator)
if isinstance(data, fluid.framework.Variable):
data = [data]
with pd.do():
ins = list(map(pd.read_input, data))
if len(ins) == 1:
ins = ins[0]
loss = generator.send(ins) # patch input
pd.write_output(loss)
loss = pd()
else:
data = next(generator)
loss = generator.send(data)
self.assertIsNotNone(loss)
avg_loss = fluid.layers.mean(loss)
fluid.backward.append_backward(loss=avg_loss)
exe = fluid.Executor(place)
exe.run(startup)
if use_gpu:
profile_type = 'GPU'
else:
profile_type = 'CPU'
with profiler.profiler(profile_type, 'total', '/tmp/profiler'):
return exe.run(main, feed=feed, fetch_list=fetch)
def _assert_same_(self, fetch, *args):
"""
Assert the return values of `run_test` are same.
Args:
fetch: Fetch list. Used for print error message
*args: The fetch result lists of each situations.
Returns:
None
Raises:
AssertionError
"""
def _impl_(a, b, fetch_id, item_id):
item_str = [
'CPU', 'ParallelCPU', 'GPU', 'ParallelGPU', 'ParallelGPUNCCL'
]
flag = numpy.allclose(a, b, rtol=0.1, atol=1e-3)
self.assertTrue(flag,
"The {0} are different in {1}, {2} vs {3}".format(
fetch[fetch_id], item_str[item_id], a, b))
for i, items in enumerate(zip(*args)):
self.assertGreater(len(items), 0)
for j in range(1, len(items)):
_impl_(items[0], items[j], fetch_id=i, item_id=j)
class ParallelOpTest(BaseParallelForTest):
@staticmethod
def __network__():
x = fluid.layers.data(shape=[784], dtype='float32', name='img')
x = yield x
hidden = fluid.layers.fc(input=x, size=200, param_attr='fc1.w')
hidden = fluid.layers.batch_norm(input=hidden)
loss = fluid.layers.mean(hidden)
yield loss
def test_simple_fc(self):
self.run_test(
callback=self.__network__,
feed={
'img': numpy.random.random(size=(51, 784)).astype('float32')
},
fetch=['fc1.w@GRAD'])
def test_fc_with_tiny_data(self):
self.run_test(
callback=self.__network__,
feed={'img': numpy.random.random(size=(1, 784)).astype('float32')},
fetch=['fc1.w@GRAD'])
class ParallelOpTestMultipleInput(BaseParallelForTest):
@staticmethod
def __network__():
x = fluid.layers.data(
shape=[784], dtype='float32', name='img1', stop_gradient=False)
y = fluid.layers.data(
shape=[784], dtype='float32', name='img2', stop_gradient=False)
yield [x, y]
x = x + y
hidden1 = fluid.layers.fc(input=x, size=200, param_attr='fc1.w')
hidden2 = fluid.layers.fc(input=hidden1, size=200, param_attr='fc2.w')
hidden3 = fluid.layers.fc(input=hidden2, size=200, param_attr='fc3.w')
loss = fluid.layers.mean(hidden3)
yield loss
def test_simple_fc(self):
self.run_test(
callback=self.__network__,
feed={
'img1': numpy.random.random(size=(51, 784)).astype('float32'),
'img2': numpy.random.random(size=(51, 784)).astype('float32')
},
fetch=['fc1.w@GRAD', 'fc2.w@GRAD', 'fc3.w@GRAD'])
if __name__ == '__main__':
unittest.main()
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import print_function
import unittest
from test_transpose_op import TestTransposeOp
class TestTransposeMKLDNN(TestTransposeOp):
def init_op_type(self):
self.op_type = "transpose2"
self.use_mkldnn = True
self.is_test = True
return
def test_check_grad(self):
return
def test_check_grad_no_input(self):
return
def test_check_grad_no_filter(self):
return
class TestCase0MKLDNN(TestTransposeMKLDNN):
def initTestCase(self):
self.shape = (3, )
self.axis = (0, )
class TestCase1a(TestTransposeMKLDNN):
def initTestCase(self):
self.shape = (3, 4, 5)
self.axis = (0, 2, 1)
class TestCase1b(TestTransposeMKLDNN):
def initTestCase(self):
self.shape = (3, 4, 5)
self.axis = (2, 1, 0)
class TestCase2(TestTransposeMKLDNN):
def initTestCase(self):
self.shape = (2, 3, 4, 5)
self.axis = (0, 2, 3, 1)
class TestCase3(TestTransposeMKLDNN):
def initTestCase(self):
self.shape = (2, 3, 4, 5, 6)
self.axis = (4, 2, 3, 1, 0)
class TestCase4(TestTransposeMKLDNN):
def initTestCase(self):
self.shape = (2, 3, 4, 5, 6, 1)
self.axis = (4, 2, 3, 1, 0, 5)
if __name__ == '__main__':
unittest.main()
...@@ -21,15 +21,24 @@ from op_test import OpTest ...@@ -21,15 +21,24 @@ from op_test import OpTest
class TestTransposeOp(OpTest): class TestTransposeOp(OpTest):
def setUp(self): def setUp(self):
self.init_op_type()
self.initTestCase() self.initTestCase()
self.op_type = "transpose2"
self.inputs = {'X': np.random.random(self.shape).astype("float32")} self.inputs = {'X': np.random.random(self.shape).astype("float32")}
self.attrs = {'axis': list(self.axis)} self.attrs = {
'axis': list(self.axis),
'use_mkldnn': self.use_mkldnn,
'is_test': self.is_test,
}
self.outputs = { self.outputs = {
'XShape': np.random.random(self.shape).astype("float32"), 'XShape': np.random.random(self.shape).astype("float32"),
'Out': self.inputs['X'].transpose(self.axis) 'Out': self.inputs['X'].transpose(self.axis)
} }
def init_op_type(self):
self.op_type = "transpose2"
self.use_mkldnn = False
self.is_test = False
def test_check_output(self): def test_check_output(self):
self.check_output(no_check_set=['XShape']) self.check_output(no_check_set=['XShape'])
......
...@@ -35,14 +35,14 @@ dtype_to_size = { ...@@ -35,14 +35,14 @@ dtype_to_size = {
} }
SUB_BLOCK_OPS = [ SUB_BLOCK_OPS = [
"while", "while_grad", "parallel_do", "parallel_do_grad", "while", "while_grad", "conditional_block", "conditional_block_grad"
"conditional_block", "conditional_block_grad"
] ]
SUB_BLOCK_PAIR = [("while", "while_grad"), ("parallel_do", "parallel_do_grad"), SUB_BLOCK_PAIR = [("while", "while_grad"),
("conditional_block", "conditional_block_grad")] ("conditional_block", "conditional_block_grad")]
PRINT_LOG = False PRINT_LOG = False
FLAGS_memory_optimize = ""
class OrderedSet(MutableSet): class OrderedSet(MutableSet):
...@@ -121,6 +121,7 @@ class ControlFlowGraph(object): ...@@ -121,6 +121,7 @@ class ControlFlowGraph(object):
self._defs = defaultdict(OrderedSet) self._defs = defaultdict(OrderedSet)
self._live_in = defaultdict(OrderedSet) self._live_in = defaultdict(OrderedSet)
self._live_out = defaultdict(OrderedSet) self._live_out = defaultdict(OrderedSet)
self._skip_opt = skip_opt self._skip_opt = skip_opt
self.pool = [] self.pool = []
...@@ -144,7 +145,6 @@ class ControlFlowGraph(object): ...@@ -144,7 +145,6 @@ class ControlFlowGraph(object):
for i in range(self.op_size): for i in range(self.op_size):
self._uses[i].update(self._ops[i].input_arg_names()) self._uses[i].update(self._ops[i].input_arg_names())
self._defs[i].update(self._ops[i].output_arg_names()) self._defs[i].update(self._ops[i].output_arg_names())
self._live_in[i] = self._uses[i]
def _update_graph(self, old_name, new_name, begin_idx=0): def _update_graph(self, old_name, new_name, begin_idx=0):
for i in range(begin_idx, self.op_size): for i in range(begin_idx, self.op_size):
...@@ -177,20 +177,52 @@ class ControlFlowGraph(object): ...@@ -177,20 +177,52 @@ class ControlFlowGraph(object):
worklist.append(d) worklist.append(d)
def _fill_pool(self, i, is_forward): def _fill_pool(self, i, is_forward):
def comparator(x, cache):
x_shape = x[1]
cache_shape = cache[1]
x_size = abs(reduce(lambda x, y: x * y, x_shape))
cache_size = abs(reduce(lambda x, y: x * y, cache_shape))
if (x_shape[0] == -1 and cache_shape[0] == -1) or \
(x_shape[0] != -1 and cache_shape[0] != -1) :
return x_size <= cache_size
else:
return False
def find_var_in_block(x):
known_vars = set()
for op in self._ops:
known_vars.update(op.output_arg_names())
return x in known_vars
block_desc = self._ops[i].block() block_desc = self._ops[i].block()
in_diff, _ = self._get_diff(self._live_in[i], self._live_out[i]) in_diff, _ = self._get_diff(self._live_in[i], self._live_out[i])
# NOTE: must sort the in_diff set for cases that get different cache var. # NOTE: must sort the in_diff set for cases that get different cache var.
# FIXME(typhoonzero): maybe use a "sorted set" is better than this. # FIXME(typhoonzero): maybe use a "sorted set" is better than this.
can_optimize = [ can_optimize = [
x for x in in_diff x for x in sorted(in_diff)
if self._check_var_validity(block_desc, x, is_forward) if self._check_var_validity(block_desc, x, is_forward)
] ]
if can_optimize: if can_optimize:
for var_name in can_optimize: for var_name in can_optimize:
cache = (var_name, self._find_var(block_desc, var_name, cache = (var_name, self._find_var(block_desc, var_name,
is_forward).shape()) is_forward).shape())
if cache not in self.pool: if cache not in self.pool and find_var_in_block(var_name):
self.pool.append(cache) i = 0
while i < len(self.pool):
mycache = self.pool[i]
mysize = mycache[1][0]
cache_size = cache[1][0]
if (mysize == -1 and cache_size == -1) or \
(mysize != -1 and cache_size != -1):
if comparator(mycache, cache):
i += 1
else:
break
elif mysize == -1 and cache_size != -1:
i += 1
elif mysize != -1 and cache_size == -1:
break
self.pool.insert(i, cache)
def _get_diff(self, a, b): def _get_diff(self, a, b):
u = a & b u = a & b
...@@ -229,7 +261,7 @@ class ControlFlowGraph(object): ...@@ -229,7 +261,7 @@ class ControlFlowGraph(object):
def _update_skip_opt_set(self): def _update_skip_opt_set(self):
for i in range(self.op_size): for i in range(self.op_size):
op = self._ops[i] op = self._ops[i]
if op.type() == "fill_constant" and op.attr("force_cpu") == True: if op.has_attr("force_cpu") and op.attr("force_cpu") == True:
self._skip_opt.update(op.output_arg_names()) self._skip_opt.update(op.output_arg_names())
def release_memory(self, skip_opt_set=None): def release_memory(self, skip_opt_set=None):
...@@ -281,6 +313,7 @@ class ControlFlowGraph(object): ...@@ -281,6 +313,7 @@ class ControlFlowGraph(object):
# update skip set to meet users' demand # update skip set to meet users' demand
if skip_opt_set: if skip_opt_set:
self._skip_opt.update(skip_opt_set) self._skip_opt.update(skip_opt_set)
counter = 0
for i in range(self.op_size): for i in range(self.op_size):
op = self._ops[i] op = self._ops[i]
if op.type() in SUB_BLOCK_OPS: if op.type() in SUB_BLOCK_OPS:
...@@ -301,6 +334,9 @@ class ControlFlowGraph(object): ...@@ -301,6 +334,9 @@ class ControlFlowGraph(object):
# If x is both in uses and defs, it can not be optimized! # If x is both in uses and defs, it can not be optimized!
if x in self._uses[i]: if x in self._uses[i]:
continue continue
if x == FLAGS_memory_optimize:
print("start match var ", x, " of op ", op.type())
print(self.pool)
for index, cache_pair in enumerate(self.pool): for index, cache_pair in enumerate(self.pool):
cache_var = cache_pair[0] cache_var = cache_pair[0]
cache_shape = cache_pair[1] cache_shape = cache_pair[1]
...@@ -323,15 +359,13 @@ class ControlFlowGraph(object): ...@@ -323,15 +359,13 @@ class ControlFlowGraph(object):
if not compare_shape(x_shape, cache_shape, level): if not compare_shape(x_shape, cache_shape, level):
continue continue
# TODO(qijun): dtype_to_size[x_dtype] and dtype_to_size[cache_dtype] # TODO(qijun): dtype_to_size[x_dtype] and dtype_to_size[cache_dtype]
if x_dtype != cache_dtype:
continue
if PRINT_LOG: if PRINT_LOG:
print(("Hit Cache !!!! cache pool index " print(
"is %d, var name is %s, " ("!!! %d, %s => %s, cache idx %d, pool size %d"
"cached var name is %s, " % (counter, x + str(x_shape),
"var shape is %s ") % (index, x, cache_var, cache_var + str(cache_shape), index,
str(cache_shape))) len(self.pool))))
counter += 1
self.pool.pop(index) self.pool.pop(index)
# Rename the var to the cache var already with # Rename the var to the cache var already with
# memory allocated in order to reuse the memory. # memory allocated in order to reuse the memory.
...@@ -484,8 +518,11 @@ def memory_optimize(input_program, ...@@ -484,8 +518,11 @@ def memory_optimize(input_program,
if level != 0 and level != 1: if level != 0 and level != 1:
raise ValueError("only support opt_level 0 or 1.") raise ValueError("only support opt_level 0 or 1.")
if skip_opt_set is not None and not isinstance(skip_opt_set, set): if skip_opt_set is not None:
raise ValueError("only support skip_opt_set as set.") if isinstance(skip_opt_set, set) or isinstance(skip_opt_set, list):
skip_opt_set = set(skip_opt_set)
else:
raise ValueError("only support skip_opt_set as set.")
global PRINT_LOG global PRINT_LOG
PRINT_LOG = print_log PRINT_LOG = print_log
if skip_grads: if skip_grads:
......
...@@ -160,10 +160,11 @@ if '${WITH_FLUID_ONLY}'== 'OFF': ...@@ -160,10 +160,11 @@ if '${WITH_FLUID_ONLY}'== 'OFF':
# put all thirdparty libraries in paddle.libs # put all thirdparty libraries in paddle.libs
libs_path='${PADDLE_BINARY_DIR}/python/paddle/libs' libs_path='${PADDLE_BINARY_DIR}/python/paddle/libs'
if os.name != 'nt':
package_data['paddle.libs']= [] package_data['paddle.libs']= []
package_data['paddle.libs']=['libwarpctc' + ext_name] package_data['paddle.libs']=[('libwarpctc' if os.name != 'nt' else 'warpctc') + ext_name]
shutil.copy('${WARPCTC_LIBRARIES}', libs_path) shutil.copy('${WARPCTC_LIBRARIES}', libs_path)
if '${WITH_MKL}' == 'ON': if '${WITH_MKL}' == 'ON':
shutil.copy('${MKLML_LIB}', libs_path) shutil.copy('${MKLML_LIB}', libs_path)
shutil.copy('${MKLML_IOMP_LIB}', libs_path) shutil.copy('${MKLML_IOMP_LIB}', libs_path)
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册