Files
paddlepaddle--paddle/paddle/cinn/optim/schedule_block_dce.h
T
2026-07-13 12:40:42 +08:00

64 lines
2.7 KiB
C++

// Copyright (c) 2024 CINN 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.
/**
* This file implements the strategy to remove the unnecessary schedule_block.
*/
#pragma once
#include <vector>
#include "paddle/cinn/common/common.h"
#include "paddle/cinn/ir/ir.h"
namespace cinn {
namespace optim {
/**
*
* This pass eliminates dead schedule blocks (those that do not affect the final
* program output).
*
* This pass is applicable in scenarios where there are redundant computations
* or store operations in the program. For example, certain schedule blocks may
* not influence the final output or may be completely unused. This commonly
* occurs in deeply nested loops where some sub-blocks are executed, but their
* results are never used, leading to wasted computation. The goal of this pass
* is to analyze the dependencies and data flow of schedule blocks to identify
* and remove those that are unnecessary, simplifying the code and improving
* execution efficiency.
*
* When applied, this pass will perform the following modifications to the IR:
* 1. The IR will be scanned for schedule blocks that do not influence the final
* output. These blocks will be marked as dead schedule blocks.
* 2. Dead schedule blocks will be removed from the IR, eliminating unnecessary
* computations. Specifically, in loops and conditional branches, any redundant
* schedule blocks will be removed.
* 3. While removing dead schedule blocks, other valid blocks in the IR will
* remain unaffected, ensuring the correctness of the program.
* 4. The final IR will be simplified, containing only the relevant computations
* that contribute to the output, resulting in more efficient execution.
*
* Performance impact: This pass addresses performance issues by removing
* unnecessary computations, potentially improving program execution efficiency,
* especially when there are large amounts of unused computations. By
* eliminating dead schedule blocks, execution time and memory consumption can
* be reduced.
*
*/
void EliminateDeadScheduleBlock(Expr* e,
const std::vector<std::string>& output_names);
} // namespace optim
} // namespace cinn