auto_parallel_recompute.py

# Copyright (c) 2021 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.

import copy
import logging

from .pass_base import PassBase, register_pass
from paddle.fluid import core, unique_name
from paddle.fluid import framework as framework
from paddle.fluid.framework import Variable, Operator
from paddle.fluid.backward import _append_grad_suffix_, _get_no_grad_set_name
from paddle.fluid.backward import ProgramStats, _rename_arg_, _find_op_path_
from paddle.distributed.auto_parallel.process_mesh import ProcessMesh
from paddle.distributed.auto_parallel.dist_attribute import OperatorDistributedAttribute
from paddle.distributed.auto_parallel.utils import get_loss_op, set_var_dist_attr, set_dist_op_desc_original_id
from paddle.distributed.auto_parallel.utils import naive_set_dist_op_attr_for_program_by_mesh_and_mapping


class RecomputeState(ProgramStats):

    def __init__(self, block, ops):
        super(RecomputeState, self).__init__(block=block, ops=ops)
        self._block = block
        self._ops = ops
        self.var_op_deps = {}

    def build_stats(self):
        for i, op in enumerate(self._ops):
            for name in op.desc.input_arg_names():
                if name in self.var_op_deps:
                    self.var_op_deps[name]["var_as_input_ops"].extend([i])
                else:
                    self.var_op_deps[name] = {}
                    self.var_op_deps[name]["var_as_input_ops"] = [i]
                    self.var_op_deps[name]["var_as_output_ops"] = []

            for name in op.desc.output_arg_names():
                if name in self.var_op_deps:
                    self.var_op_deps[name]["var_as_output_ops"].extend([i])
                else:
                    self.var_op_deps[name] = {}
                    self.var_op_deps[name]["var_as_input_ops"] = []
                    self.var_op_deps[name]["var_as_output_ops"] = [i]

    def get_recompute_segments(self, checkpoints):
        """ get recompute segments from checkpoints """
        segments = []
        start_idx = -1
        pre_segment_end_idx = -1
        while start_idx + 1 < len(checkpoints):
            if start_idx == -1:
                ckpt_name = checkpoints[start_idx + 1]
                if ckpt_name not in self.var_op_deps:
                    start_idx += 1
                    continue
                op_idx_list = self.var_op_deps[ckpt_name]["var_as_output_ops"]
                if op_idx_list:
                    segments.append([0, max(op_idx_list) + 1])
            else:
                flag, min_idx, max_idx = self.is_subgraph(
                    [checkpoints[start_idx]], [checkpoints[start_idx + 1]])
                if flag:
                    min_idx = self._update_segment_start(
                        min_idx, pre_segment_end_idx)
                    segments.append([min_idx, max_idx + 1])
                else:
                    logging.info(
                        "Could not recompute op range [{}] - [{}] ".format(
                            min_idx, max_idx + 1))
            start_idx += 1

        for i, (idx1, idx2) in enumerate(segments):
            logging.info("recompute segment[{}]".format(i))
            logging.info("segment start op: [{}]: [{}] [{}]".format(
                self._ops[idx1].desc.type(),
                self._ops[idx1].desc.input_arg_names(),
                self._ops[idx1].desc.output_arg_names()))
            logging.info("segment end op: [{}]: [{}] [{}]".format(
                self._ops[idx2 - 1].desc.type(),
                self._ops[idx2 - 1].desc.input_arg_names(),
                self._ops[idx2 - 1].desc.output_arg_names()))

        return segments

    def modify_forward_desc_for_recompute(self, dist_context):
        """
        If program's foward part has 'dropout' op, this function will insert 
        a seed op before it to guarantee that two dropout op have the same outputs.
        """
        op_types = [op.desc.type() for op in self._ops]
        if "dropout" not in op_types:
            return

        op_idx = 0
        while op_idx < len(self._ops):
            cur_op = self._ops[op_idx]
            if "grad" in cur_op.type:
                break
            if cur_op.type != "dropout":
                op_idx += 1
                continue
            if cur_op.input("Seed") is not None and len(cur_op.input("Seed")):
                op_idx += 1
                continue

            cur_op_dist_attr = dist_context.get_op_dist_attr_for_program(cur_op)
            # insert seed op to guarantee that two dropout op have the same outputs
            op_unique_name = unique_name.generate("seed")
            var_unique_name = unique_name.generate_with_ignorable_key(".".join(
                [op_unique_name, 'tmp']))
            seed_var = self._block.create_var(
                name=var_unique_name,
                dtype='int32',
                type=core.VarDesc.VarType.LOD_TENSOR,
                persistable=False,
                stop_gradient=False)

            # set new seed_var's dist_attr
            ref_dims_mapping = [-1]
            ref_process_mesh = cur_op_dist_attr.process_mesh
            seed_var_dist_attr = set_var_dist_attr(dist_context, seed_var,
                                                   ref_dims_mapping,
                                                   ref_process_mesh)

            seed = 0 if cur_op.attr("fix_seed") is False else int(
                cur_op.attr("seed"))
            seed_op = self._block._insert_op_without_sync(
                index=cur_op.idx,
                type="seed",
                inputs={},
                outputs={"Out": seed_var},
                attrs={
                    "seed": seed,
                    "force_cpu": True
                })
            # set new seed op's dist_attr
            naive_set_dist_op_attr_for_program_by_mesh_and_mapping(
                seed_op, ref_process_mesh, ref_dims_mapping, dist_context)

            # modify dropout op's desc
            self._ops.insert(op_idx, seed_op)
            cur_op.desc.set_input("Seed", [var_unique_name])
            cur_op.desc.remove_attr("fix_seed")
            cur_op.desc.remove_attr("seed")
            cur_op_dist_attr.set_input_dist_attr(seed_var.name,
                                                 seed_var_dist_attr)
            self._block._sync_with_cpp()
            op_idx += 2


def _find_op_index(block, cur_op):
    for idx in range(block.desc.op_size()):
        if cur_op.desc == block.desc.op(idx):
            return idx
    return -1


def _get_stop_gradients(program, no_grad_set):
    """ get no grad var """
    if no_grad_set is None:
        no_grad_set = set()
    else:
        no_grad_set = _get_no_grad_set_name(no_grad_set)

    no_grad_set_name = set()
    for var in program.list_vars():
        assert isinstance(var, Variable)
        if "@GRAD" in var.name:
            break
        if var.stop_gradient:
            no_grad_set_name.add(_append_grad_suffix_(var.name))
    no_grad_set_name.update(list(map(_append_grad_suffix_, no_grad_set)))
    return no_grad_set_name


def _add_needed_descs_to_block(descs, block, main_block, in_memory_vars,
                               dist_context):
    """
    Get the recomputed ops which will insert the backward part
    """
    if len(descs) == 0:
        return []
    result_descs = []
    op_role_attr_name = core.op_proto_and_checker_maker.kOpRoleAttrName()
    backward = core.op_proto_and_checker_maker.OpRole.Backward
    for desc in descs:
        if isinstance(desc, framework.Operator):
            desc = desc.desc
        if isinstance(desc, tuple):
            desc = desc[0]
        is_needed = False
        for name in desc.output_arg_names():
            if main_block.has_var(name) and main_block.var(name).persistable:
                continue
            if name not in in_memory_vars:
                is_needed = True
        if is_needed:
            new_op_desc = block.desc.append_op()
            new_op_desc.copy_from(desc)
            set_dist_op_desc_original_id(new_op_desc, desc, dist_context)
            new_op_desc._set_attr(op_role_attr_name, backward)
            result_descs.append(new_op_desc)
    return result_descs


@register_pass("auto_parallel_recompute")
class RecomputePass(PassBase):

    def __init__(self):
        super(RecomputePass, self).__init__()
        self.set_attr("checkpoints", None)
        self.set_attr("loss", None)
        self.set_attr("dist_context", None)
        self.set_attr("no_grad_set", None)

    def _check_self(self):
        if self.get_attr("dist_context") is None:
            return False
        if self.get_attr("loss") is None:
            return False
        if self.get_attr("checkpoints") is None:
            return False
        return True

    def _check_conflict(self, other_pass):
        return True

    def _apply_single_impl(self, main_programs, startup_programs, context):
        checkpoints = self.get_attr("checkpoints")
        loss = self.get_attr("loss")
        no_grad_set = self.get_attr("no_grad_set")
        self._dist_context = self.get_attr("dist_context")

        main_block = main_programs.global_block()
        no_grad_set_name = _get_stop_gradients(main_programs, no_grad_set)
        # get op_path which is related to loss
        op_path = _find_op_path_(main_block, [loss], [], no_grad_set_name)

        # step 1: build recompute state
        rc_state = RecomputeState(main_block, op_path)
        rc_state.modify_forward_desc_for_recompute(self._dist_context)
        rc_state.build_stats()
        checkpoints = rc_state.sort_checkpoints(checkpoints)
        segments = rc_state.get_recompute_segments(checkpoints)
        if segments == []:
            return

        # step 2: get vars_should_be_hold
        vars_should_be_hold = []
        for segment in segments:
            vars_should_be_hold.extend(
                rc_state.get_out_of_subgraph_vars(segment[0], segment[1]))
        cross_vars = set(vars_should_be_hold) - set(checkpoints)
        logging.info(
            "found [{}] vars which cross recompute segment: [{}],"
            "better checkpoints might be set to reduce those vars".format(
                len(cross_vars), cross_vars))
        vars_should_be_hold.extend(rc_state.get_reserved_vars())
        vars_should_be_hold.extend(rc_state.get_input_nodes())
        vars_should_be_hold = list(set(vars_should_be_hold))
        vars_in_memory = vars_should_be_hold + checkpoints

        # step 3: get recomputed fwd ops desc
        var_name_dict = {}
        ckpt_ops_dict = {}
        buffer_block = main_block.program._create_block()
        for i, segment in enumerate(segments[::-1]):
            fwd_ops = op_path[segment[0]:segment[1]]
            var_suffix = ".subprog_%d" % i
            for op in fwd_ops:
                input_and_output_names = []
                input_and_output_names.extend(op.desc.input_arg_names())
                input_and_output_names.extend(op.desc.output_arg_names())
                cur_op_dist_attr = self._dist_context.get_op_dist_attr_for_program(
                    op)
                assert cur_op_dist_attr is not None
                for name in input_and_output_names:
                    if main_block.var(name).persistable or name in checkpoints:
                        continue
                    if name in vars_should_be_hold:
                        continue
                    if name not in var_name_dict:
                        ref_process_mesh = cur_op_dist_attr.process_mesh
                        if name in op.desc.input_arg_names():
                            ref_dims_mapping = cur_op_dist_attr.get_input_dims_mapping(
                                name)
                        else:
                            ref_dims_mapping = cur_op_dist_attr.get_output_dims_mapping(
                                name)
                        # record recomputed var's old_name and new_name (old_name.subprog_XXX)
                        # create new var with new name
                        var_name_dict[name] = name + var_suffix
                        ref_var = main_block.var(name)
                        rc_var = main_block.create_var(
                            name=var_name_dict[name],
                            shape=ref_var.shape,
                            dtype=ref_var.dtype,
                            type=ref_var.type,
                            persistable=ref_var.persistable,
                            stop_gradient=ref_var.stop_gradient)
                        # set new recomputed var's dist attr
                        set_var_dist_attr(self._dist_context, rc_var,
                                          ref_dims_mapping, ref_process_mesh)
            # get recomputed segment's descs
            segment_descs = _add_needed_descs_to_block(fwd_ops, buffer_block,
                                                       main_block,
                                                       vars_in_memory,
                                                       self._dist_context)
            # rename recomputed ops' input and output var name
            for key in var_name_dict:
                _rename_arg_(segment_descs, key, var_name_dict[key])

            # NOTE: one forward op could be correspond to multiple xxx_grad op.
            # When traversing all grad_ops in reverse, need to set a flag to indicate
            # whether the ckpt and its segment_descs can be used.
            ckpt_op = op_path[segment[1] - 1]
            ckpt_ops_dict[ckpt_op.desc.original_id()] = [True, segment_descs]

        # step 4: insert recomputed fwd ops
        ops = main_block.ops
        loss_op = get_loss_op(main_block)
        loss_op_idx = _find_op_index(main_block, loss_op)
        dist_op_context = self._dist_context.dist_op_context
        assert loss_op_idx != -1
        # Traversing all grad_ops in reverse, and if the fwd op corresponding to reverse op is checkpoints,
        # segments ops should be inserted.
        for i in range(len(ops) - 1, loss_op_idx, -1):
            grad_op = ops[i]
            # remove some attrs of dropout_grad op's desc
            if grad_op.type == "dropout_grad":
                grad_op.desc.remove_attr("fix_seed")
                grad_op.desc.remove_attr("seed")
                main_block._sync_with_cpp()

            # rename grad op's var_name which is not in 'vars_in_memory'
            for key in var_name_dict:
                self.reset_op_dist_attr(grad_op, var_name_dict)
                _rename_arg_([grad_op.desc], key, var_name_dict[key])

            # insert recomputed ops
            original_id = grad_op.desc.original_id()
            if original_id in dist_op_context.grad_op_id_to_op_id:
                fwd_op_id = dist_op_context.grad_op_id_to_op_id[original_id]
                if fwd_op_id in ckpt_ops_dict and ckpt_ops_dict[fwd_op_id][0]:
                    idx = grad_op.idx
                    while idx - 1 >= 0 and ops[idx - 1].type == "sum":
                        idx -= 1
                    segment_descs = ckpt_ops_dict[fwd_op_id][1]
                    for _, op_desc in reversed(list(enumerate(segment_descs))):
                        rc_desc = main_block.desc._insert_op(idx)
                        rc_desc.copy_from(op_desc)
                        rc_desc.set_original_id(rc_desc.id())
                        rc_op = Operator(main_block, rc_desc)
                        main_block.ops.insert(idx, rc_op)
                        # set recomputed ops' dist attr
                        fwd_op_dist_attr = self._dist_context.get_op_dist_attr_for_program_with_id(
                            op_desc.original_id())
                        assert fwd_op_dist_attr is not None
                        self.set_op_dist_attr(rc_op, fwd_op_dist_attr,
                                              var_name_dict)

                    ckpt_ops_dict[fwd_op_id][0] = False
                    main_block._sync_with_cpp()

        main_programs._sync_with_cpp()

    def reset_op_dist_attr(self, op, var_name_dict):
        op_dist_attr = self._dist_context.get_op_dist_attr_for_program(op)
        assert op_dist_attr is not None
        for input in op.desc.input_arg_names():
            if input in var_name_dict.keys():
                in_dist_attr = op_dist_attr.get_input_dist_attr(input)
                op_dist_attr.set_input_dist_attr(var_name_dict[input],
                                                 in_dist_attr)
        for output in op.desc.output_arg_names():
            if output in var_name_dict.keys():
                out_dist_attr = op_dist_attr.get_output_dist_attr(output)
                op_dist_attr.set_output_dist_attr(var_name_dict[output],
                                                  out_dist_attr)

    def set_op_dist_attr(self, op, old_dist_attr, var_name_dict):
        new_dist_attr = OperatorDistributedAttribute()
        new_dist_attr.is_recompute = True
        new_dist_attr.impl_idx = old_dist_attr.impl_idx
        new_dist_attr.impl_type = old_dist_attr.impl_type
        new_dist_attr.process_mesh = old_dist_attr.process_mesh
        for input in old_dist_attr.inputs_dist_attrs.keys():
            if input in var_name_dict.keys():
                in_dist_attr = old_dist_attr.inputs_dist_attrs[input]
                new_dist_attr.set_input_dist_attr(var_name_dict[input],
                                                  in_dist_attr)
            else:
                in_dist_attr = old_dist_attr.inputs_dist_attrs[input]
                new_dist_attr.set_input_dist_attr(input, in_dist_attr)
        for output in old_dist_attr.outputs_dist_attrs.keys():
            if output in var_name_dict.keys():
                out_dist_attr = old_dist_attr.outputs_dist_attrs[output]
                new_dist_attr.set_output_dist_attr(var_name_dict[output],
                                                   out_dist_attr)
            else:
                out_dist_attr = old_dist_attr.outputs_dist_attrs[output]
                new_dist_attr.set_output_dist_attr(output, out_dist_attr)
        self._dist_context.set_op_dist_attr_for_program(op, new_dist_attr)