* Copyright (c) 2025 Huawei Technologies Co., Ltd.
* This program is free software, you can redistribute it and/or modify it under the terms and conditions of
* CANN Open Software License Agreement Version 2.0 (the "License").
* Please refer to the License for details. You may not use this file except in compliance with the License.
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
* INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
* See LICENSE in the root of the software repository for the full text of the License.
*/
* \file test_expand_operation.cpp
* \brief
*/
#include "test_operation.h"
using namespace tile_fwk::test_operation;
namespace {
struct ExpandOpFuncArgs : public OpFuncArgs {
ExpandOpFuncArgs(const std::vector<int64_t>& viewShape, const std::vector<int64_t> tileShape)
: viewShape_(viewShape), tileShape_(tileShape)
{}
std::vector<int64_t> viewShape_;
std::vector<int64_t> tileShape_;
};
struct ExpandOpMetaData {
explicit ExpandOpMetaData(const OpFunc& opFunc, const nlohmann::json& test_data)
: opFunc_(opFunc), test_data_(test_data)
{}
OpFunc opFunc_;
nlohmann::json test_data_;
};
void UpdateInputExpandViewShape(
std::vector<int64_t>& inputViewShape, const std::vector<SymbolicScalar>& inputsShape,
const std::vector<SymbolicScalar>& outputsShape)
{
for (size_t i = 0; i < inputsShape.size(); i++) {
if (inputsShape[i] == 1 && outputsShape[i] != 1) {
inputViewShape[i] = 1;
}
}
}
void UpdateInputExpandVaildShape(
std::vector<SymbolicScalar>& inputValidShape, const std::vector<SymbolicScalar>& inputsShape,
const std::vector<SymbolicScalar>& outputsShape)
{
for (size_t i = 0; i < inputsShape.size(); i++) {
if (inputsShape[i] == 1 && outputsShape[i] != 1) {
inputValidShape[i] = 1;
}
}
}
void UpdateInputExpandOffset(
std::vector<SymbolicScalar>& inputOffset, const std::vector<SymbolicScalar>& inputsShape,
const std::vector<SymbolicScalar>& outputsShape)
{
for (size_t i = 0; i < inputsShape.size(); i++) {
if (inputsShape[i] == 1 && outputsShape[i] != 1) {
inputOffset[i] = 0;
}
}
}
static void ExpandOperationExeFunc2Dims(
const std::vector<Tensor>& inputs, std::vector<Tensor>& outputs, const OpFuncArgs* opArgs)
{
FUNCTION("main", {inputs[0]}, {outputs[0]})
{
std::vector<SymbolicScalar> inputsShape = {inputs[0].GetShape()[0], inputs[0].GetShape()[1]};
std::vector<SymbolicScalar> outputsShape = {outputs[0].GetShape()[0], outputs[0].GetShape()[1]};
auto args = static_cast<const ExpandOpFuncArgs*>(opArgs);
std::vector<int64_t> viewShape = {args->viewShape_[0], args->viewShape_[1]};
std::vector<int64_t> inputViewShape = viewShape;
UpdateInputExpandViewShape(inputViewShape, inputsShape, outputsShape);
std::vector<SymbolicScalar> inputValidShape(2, 0);
std::vector<SymbolicScalar> inputOffset(2, 0);
const int bloop = CeilDiv(outputsShape[0], viewShape[0]);
const int sloop = CeilDiv(outputsShape[1], viewShape[1]);
LOOP("LOOP_L0_bIdx", FunctionType::DYNAMIC_LOOP, bIdx, LoopRange(0, bloop, 1))
{
LOOP("LOOP_L1_sIdx", FunctionType::DYNAMIC_LOOP, sIdx, LoopRange(0, sloop, 1))
{
inputValidShape = {
std::min(inputsShape[0] - bIdx * inputViewShape[0], inputViewShape[0]),
std::min(inputsShape[1] - sIdx * inputViewShape[1], inputViewShape[1])};
inputOffset = {bIdx * inputViewShape[0], sIdx * inputViewShape[1]};
UpdateInputExpandVaildShape(inputValidShape, inputsShape, outputsShape);
UpdateInputExpandOffset(inputOffset, inputsShape, outputsShape);
Tensor tileTensor0 = View(inputs[0], inputViewShape, inputValidShape, inputOffset);
TileShape::Current().SetVecTile(args->tileShape_);
auto res = Expand(
tileTensor0, viewShape,
{std::min(outputsShape[0] - bIdx * viewShape[0], viewShape[0]),
std::min(outputsShape[1] - sIdx * viewShape[1], viewShape[1])});
Assemble(res, {bIdx * viewShape[0], sIdx * viewShape[1]}, outputs[0]);
}
}
}
}
static void ExpandOperationExeFunc3Dims(
const std::vector<Tensor>& inputs, std::vector<Tensor>& outputs, const OpFuncArgs* opArgs)
{
FUNCTION("main", {inputs[0]}, {outputs[0]})
{
std::vector<SymbolicScalar> inputsShape = {
inputs[0].GetShape()[0], inputs[0].GetShape()[1], inputs[0].GetShape()[2]};
std::vector<SymbolicScalar> outputsShape = {
outputs[0].GetShape()[0], outputs[0].GetShape()[1], outputs[0].GetShape()[2]};
auto args = static_cast<const ExpandOpFuncArgs*>(opArgs);
std::vector<int64_t> viewShape = {args->viewShape_[0], args->viewShape_[1], args->viewShape_[2]};
std::vector<int64_t> inputViewShape = viewShape;
UpdateInputExpandViewShape(inputViewShape, inputsShape, outputsShape);
std::vector<SymbolicScalar> inputValidShape(3, 0);
std::vector<SymbolicScalar> inputOffset(3, 0);
const int bloop = CeilDiv(outputsShape[0], viewShape[0]);
const int sloop = CeilDiv(outputsShape[1], viewShape[1]);
const int nloop = CeilDiv(outputsShape[2], viewShape[2]);
LOOP("LOOP_L0_bIdx", FunctionType::DYNAMIC_LOOP, bIdx, LoopRange(0, bloop, 1))
{
LOOP("LOOP_L1_sIdx", FunctionType::DYNAMIC_LOOP, sIdx, LoopRange(0, sloop, 1))
{
LOOP("LOOP_L2_nIdx", FunctionType::DYNAMIC_LOOP, nIdx, LoopRange(0, nloop, 1))
{
inputValidShape = {
std::min(inputsShape[0] - bIdx * inputViewShape[0], inputViewShape[0]),
std::min(inputsShape[1] - sIdx * inputViewShape[1], inputViewShape[1]),
std::min(inputsShape[2] - nIdx * inputViewShape[2], inputViewShape[2])};
inputOffset = {bIdx * inputViewShape[0], sIdx * inputViewShape[1], nIdx * inputViewShape[2]};
UpdateInputExpandVaildShape(inputValidShape, inputsShape, outputsShape);
UpdateInputExpandOffset(inputOffset, inputsShape, outputsShape);
Tensor tileTensor0 = View(inputs[0], inputViewShape, inputValidShape, inputOffset);
TileShape::Current().SetVecTile(args->tileShape_);
auto res = Expand(
tileTensor0, viewShape,
{std::min(outputsShape[0] - bIdx * viewShape[0], viewShape[0]),
std::min(outputsShape[1] - sIdx * viewShape[1], viewShape[1]),
std::min(outputsShape[2] - nIdx * viewShape[2], viewShape[2])});
Assemble(res, {bIdx * viewShape[0], sIdx * viewShape[1], nIdx * viewShape[2]}, outputs[0]);
}
}
}
}
}
static void ExpandOperationExeFunc4Dims(
const std::vector<Tensor>& inputs, std::vector<Tensor>& outputs, const OpFuncArgs* opArgs)
{
FUNCTION("main", {inputs[0]}, {outputs[0]})
{
std::vector<SymbolicScalar> inputsShape = {
inputs[0].GetShape()[0], inputs[0].GetShape()[1], inputs[0].GetShape()[2], inputs[0].GetShape()[3]};
std::vector<SymbolicScalar> outputsShape = {
outputs[0].GetShape()[0], outputs[0].GetShape()[1], outputs[0].GetShape()[2], outputs[0].GetShape()[3]};
auto args = static_cast<const ExpandOpFuncArgs*>(opArgs);
std::vector<int64_t> viewShape = {
args->viewShape_[0], args->viewShape_[1], args->viewShape_[2], args->viewShape_[3]};
std::vector<int64_t> inputViewShape = viewShape;
UpdateInputExpandViewShape(inputViewShape, inputsShape, outputsShape);
std::vector<SymbolicScalar> inputValidShape(4, 0);
std::vector<SymbolicScalar> inputOffset(4, 0);
const int bloop = CeilDiv(outputsShape[0], viewShape[0]);
const int sloop = CeilDiv(outputsShape[1], viewShape[1]);
const int nloop = CeilDiv(outputsShape[2], viewShape[2]);
const int mloop = CeilDiv(outputsShape[3], viewShape[3]);
LOOP("LOOP_L0_bIdx", FunctionType::DYNAMIC_LOOP, bIdx, LoopRange(0, bloop, 1))
{
LOOP("LOOP_L1_sIdx", FunctionType::DYNAMIC_LOOP, sIdx, LoopRange(0, sloop, 1))
{
LOOP("LOOP_L2_nIdx", FunctionType::DYNAMIC_LOOP, nIdx, LoopRange(0, nloop, 1))
{
LOOP("LOOP_L3_mIdx", FunctionType::DYNAMIC_LOOP, mIdx, LoopRange(0, mloop, 1))
{
inputValidShape = {
std::min(inputsShape[0] - bIdx * inputViewShape[0], inputViewShape[0]),
std::min(inputsShape[1] - sIdx * inputViewShape[1], inputViewShape[1]),
std::min(inputsShape[2] - nIdx * inputViewShape[2], inputViewShape[2]),
std::min(inputsShape[3] - mIdx * inputViewShape[3], inputViewShape[3])};
inputOffset = {
bIdx * inputViewShape[0], sIdx * inputViewShape[1], nIdx * inputViewShape[2],
mIdx * inputViewShape[3]};
UpdateInputExpandVaildShape(inputValidShape, inputsShape, outputsShape);
UpdateInputExpandOffset(inputOffset, inputsShape, outputsShape);
Tensor tileTensor0 = View(inputs[0], inputViewShape, inputValidShape, inputOffset);
TileShape::Current().SetVecTile(args->tileShape_);
auto res = Expand(
tileTensor0, viewShape,
{std::min(outputsShape[0] - bIdx * viewShape[0], viewShape[0]),
std::min(outputsShape[1] - sIdx * viewShape[1], viewShape[1]),
std::min(outputsShape[2] - nIdx * viewShape[2], viewShape[2]),
std::min(outputsShape[3] - mIdx * viewShape[3], viewShape[3])});
Assemble(
res, {bIdx * viewShape[0], sIdx * viewShape[1], nIdx * viewShape[2], mIdx * viewShape[3]},
outputs[0]);
}
}
}
}
}
}
class ExpandOperationTest : public npu::tile_fwk::stest::TestSuite_STest_Ops_Aihac_param<ExpandOpMetaData> {};
INSTANTIATE_TEST_SUITE_P(
TestExpand, ExpandOperationTest,
::testing::ValuesIn(GetOpMetaData<ExpandOpMetaData>(
{ExpandOperationExeFunc2Dims, ExpandOperationExeFunc3Dims, ExpandOperationExeFunc4Dims}, "Expand")));
TEST_P(ExpandOperationTest, TestExpand)
{
auto test_data = GetParam().test_data_;
auto args = ExpandOpFuncArgs(GetViewShape(test_data), GetTileShape(test_data));
auto testCase = CreateTestCaseDesc<ExpandOpMetaData>(GetParam(), &args);
TestExecutor::runTest(testCase);
}
}
