未验证 提交 230c6ce1 编写于 作者: Y yangguohao 提交者: GitHub

FLUID: move limit_by_capacity to PHI (#55948)

上级 81ccd99e
...@@ -12,7 +12,13 @@ ...@@ -12,7 +12,13 @@
// See the License for the specific language governing permissions and // See the License for the specific language governing permissions and
// limitations under the License. // limitations under the License.
#include "paddle/fluid/operators/limit_by_capacity_op.h" #include "paddle/fluid/framework/data_type.h"
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/framework/op_registry.h"
#if defined(PADDLE_WITH_GLOO)
#include "paddle/fluid/framework/fleet/gloo_wrapper.h"
#endif
namespace paddle { namespace paddle {
namespace operators { namespace operators {
...@@ -80,10 +86,3 @@ namespace plat = paddle::platform; ...@@ -80,10 +86,3 @@ namespace plat = paddle::platform;
REGISTER_OP_WITHOUT_GRADIENT(limit_by_capacity, REGISTER_OP_WITHOUT_GRADIENT(limit_by_capacity,
ops::LimitByCapacityOp, ops::LimitByCapacityOp,
ops::LimitByCapacityOpMaker); ops::LimitByCapacityOpMaker);
PD_REGISTER_STRUCT_KERNEL(limit_by_capacity,
CPU,
ALL_LAYOUT,
ops::LimitByCapacityOpCPUKernel,
int,
int64_t) {}
// Copyright (c) 2023 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/phi/kernels/limit_by_capacity_kernel.h"
#include "paddle/phi/core/errors.h"
#include "paddle/phi/core/kernel_registry.h"
#if defined(PADDLE_WITH_GLOO)
#include "paddle/phi/core/distributed/gloo_comm_context.h"
#endif
namespace phi {
template <typename T, typename Context>
void LimitByCapacityKernel(const Context& dev_ctx,
const DenseTensor& expert_count,
const DenseTensor& capacity,
int n_worker,
DenseTensor* Out) {
PADDLE_THROW(
phi::errors::Unimplemented("limit_by_capacity is not supported on CPU."));
}
} // namespace phi
PD_REGISTER_KERNEL(limit_by_capacity,
CPU,
ALL_LAYOUT,
phi::LimitByCapacityKernel,
int,
int64_t) {}
// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved. // Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
// //
// Licensed under the Apache License, Version 2.0 (the "License"); // Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License. // you may not use this file except in compliance with the License.
...@@ -11,22 +11,14 @@ ...@@ -11,22 +11,14 @@
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and // See the License for the specific language governing permissions and
// limitations under the License. // limitations under the License.
//
// The file has been adapted from the two files:
// https://github.com/laekov/fastmoe/blob/master/cuda/balancing.cu
// https://github.com/laekov/fastmoe/blob/master/cuda/balancing.cuh
// Git commit hash: 295a615aacce7e54a37e7935274ba15e901c78e4
// We retain the following license from the original files:
// Copyright 2021, Jiaao He. All rights reserved.
// Licensed under the Apache License, Version 2.0 (the "License").
#include "paddle/fluid/operators/limit_by_capacity_op.h" #include "paddle/phi/kernels/limit_by_capacity_kernel.h"
#include "paddle/fluid/framework/op_registry.h"
#include "paddle/fluid/platform/float16.h"
#include "paddle/phi/backends/gpu/gpu_primitives.h" #include "paddle/phi/backends/gpu/gpu_primitives.h"
#include "paddle/phi/core/dense_tensor.h"
#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/core/tensor_utils.h"
namespace paddle { namespace phi {
namespace operators {
template <typename T> template <typename T>
__global__ void limit_by_capacity_impl( __global__ void limit_by_capacity_impl(
...@@ -47,39 +39,29 @@ __global__ void limit_by_capacity_impl( ...@@ -47,39 +39,29 @@ __global__ void limit_by_capacity_impl(
} }
} }
template <typename T, typename DeviceContext> template <typename T, typename Context>
class LimitByCapacityOpCUDAKernel : public framework::OpKernel<T> { void LimitByCapacityKernel(const Context& dev_ctx,
public: const DenseTensor& expert_count,
void Compute(const framework::ExecutionContext& context) const override { const DenseTensor& capacity,
auto expert_count = context.Input<phi::DenseTensor>("expert_count"); int n_worker,
auto capacity = context.Input<phi::DenseTensor>("capacity"); DenseTensor* Out) {
auto n_worker = context.Attr<int>("n_worker"); auto expert_count_ptr = &expert_count;
auto out = context.Output<phi::DenseTensor>("Out"); auto n_expert = expert_count_ptr->numel() / n_worker;
auto n_expert = expert_count->numel() / n_worker;
const auto place = context.GetPlace();
const auto& dev_ctx = context.template device_context<phi::GPUContext>();
dim3 grid_dim(256); dim3 grid_dim(256);
dim3 block_dim(1024); dim3 block_dim(1024);
auto out_data = out->mutable_data<T>(place); auto out_data = dev_ctx.template Alloc<T>(Out);
const T* ec_data = expert_count->data<T>(); const T* ec_data = expert_count_ptr->data<T>();
phi::DenseTensor capacity_copy; phi::DenseTensor capacity_copy;
framework::TensorCopy(*capacity, place, dev_ctx, &capacity_copy); phi::Copy(dev_ctx, capacity, dev_ctx.GetPlace(), false, &capacity_copy);
T* cap_data = capacity_copy.mutable_data<T>(place); T* cap_data = dev_ctx.template Alloc<T>(&capacity_copy);
limit_by_capacity_impl<T><<<grid_dim, block_dim, 0, dev_ctx.stream()>>>( limit_by_capacity_impl<T><<<grid_dim, block_dim, 0, dev_ctx.stream()>>>(
ec_data, cap_data, out_data, n_expert, n_worker); ec_data, cap_data, out_data, n_expert, n_worker);
} }
};
} // namespace operators } // namespace phi
} // namespace paddle
namespace ops = paddle::operators; PD_REGISTER_KERNEL(
PD_REGISTER_STRUCT_KERNEL(limit_by_capacity, limit_by_capacity, GPU, ALL_LAYOUT, phi::LimitByCapacityKernel, int64_t) {}
GPU,
ALL_LAYOUT,
ops::LimitByCapacityOpCUDAKernel,
int64_t) {}
// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved. // Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
// //
// Licensed under the Apache License, Version 2.0 (the "License"); // Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License. // you may not use this file except in compliance with the License.
...@@ -13,25 +13,16 @@ ...@@ -13,25 +13,16 @@
// limitations under the License. // limitations under the License.
#pragma once #pragma once
#include "paddle/fluid/framework/data_type.h"
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/framework/op_registry.h"
#if defined(PADDLE_WITH_GLOO) #include "paddle/phi/core/dense_tensor.h"
#include "paddle/fluid/framework/fleet/gloo_wrapper.h"
#endif
namespace paddle { namespace phi {
namespace operators {
template <typename T, typename DeviceContext> template <typename T, typename Context>
class LimitByCapacityOpCPUKernel : public framework::OpKernel<T> { void LimitByCapacityKernel(const Context& dev_ctx,
public: const DenseTensor& expert_count,
void Compute(const framework::ExecutionContext& ctx) const override { const DenseTensor& capacity,
PADDLE_THROW(platform::errors::Unavailable( int n_worker,
"Do not support limit by capacity op for cpu kernel now.")); DenseTensor* Out);
}
};
} // namespace operators } // namespace phi
} // namespace paddle
// Copyright (c) 2023 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/phi/core/compat/op_utils.h"
namespace phi {
KernelSignature LimitByCapacityOpArgumentMapping(
const ArgumentMappingContext& ctx UNUSED) {
return KernelSignature(
"limit_by_capacity", {"expert_count", "capacity"}, {"n_worker"}, {"Out"});
}
} // namespace phi
PD_REGISTER_ARG_MAPPING_FN(limit_by_capacity,
phi::LimitByCapacityOpArgumentMapping);
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