derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
222 lines
11 KiB
C++
222 lines
11 KiB
C++
/*******************************************************************************
|
|
*
|
|
* MIT License
|
|
*
|
|
* Copyright (C) 2019-2022 Advanced Micro Devices, Inc. All rights reserved.
|
|
*
|
|
* Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
* of this software and associated documentation files (the "Software"), to deal
|
|
* in the Software without restriction, including without limitation the rights
|
|
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
* copies of the Software, and to permit persons to whom the Software is
|
|
* furnished to do so, subject to the following conditions:
|
|
*
|
|
* The above copyright notice and this permission notice shall be included in
|
|
* all copies or substantial portions of the Software.
|
|
*
|
|
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
* SOFTWARE.
|
|
*
|
|
*******************************************************************************/
|
|
|
|
#include "ClientProblemFactory.hpp"
|
|
#include "DataInitialization.hpp"
|
|
|
|
#include <cstddef>
|
|
|
|
namespace Tensile
|
|
{
|
|
namespace Client
|
|
{
|
|
ClientProblemFactory::ClientProblemFactory(po::variables_map const& args)
|
|
: m_freeIndices(args["free"].as<ContractionProblem::FreeIndices>())
|
|
, m_batchIndices(args["batch"].as<ContractionProblem::BatchIndices>())
|
|
, m_boundIndices(args["bound"].as<ContractionProblem::BoundIndices>())
|
|
, m_problemSizes(args["problem-size"].as<std::vector<std::vector<size_t>>>())
|
|
, m_aZeroPads(args["a-zero-pads"].as<std::vector<std::vector<size_t>>>())
|
|
, m_bZeroPads(args["b-zero-pads"].as<std::vector<std::vector<size_t>>>())
|
|
, m_aType(DataType::Float)
|
|
, m_bType(DataType::Float)
|
|
, m_cType(DataType::Float)
|
|
, m_dType(DataType::Float)
|
|
, m_alphaType(DataType::Float)
|
|
, m_betaType(DataType::Float)
|
|
, m_stridedBatched(args["strided-batched"].as<bool>())
|
|
, m_highPrecisionAccumulate(args["high-precision-accumulate"].as<bool>())
|
|
, m_kernelLanguage(args["kernel-language"].as<KernelLanguage>())
|
|
, m_performanceMetric(args["performance-metric"].as<PerformanceMetric>())
|
|
, m_deterministicMode(args["deterministic-mode"].as<bool>())
|
|
, m_cEqualsD(args["c-equal-d"].as<bool>())
|
|
, m_arithmeticUnit(args["arithmetic-unit"].as<ArithmeticUnit>())
|
|
, m_aStrides(args["a-strides"].as<std::vector<std::vector<size_t>>>())
|
|
, m_bStrides(args["b-strides"].as<std::vector<std::vector<size_t>>>())
|
|
, m_cStrides(args["c-strides"].as<std::vector<std::vector<size_t>>>())
|
|
, m_dStrides(args["d-strides"].as<std::vector<std::vector<size_t>>>())
|
|
, m_aOps(args["a-ops"].as<TensorOps>())
|
|
, m_bOps(args["b-ops"].as<TensorOps>())
|
|
, m_cOps(args["c-ops"].as<TensorOps>())
|
|
, m_dOps(args["d-ops"].as<TensorOps>())
|
|
, m_aOffset(args["offset-a"].as<size_t>())
|
|
, m_bOffset(args["offset-b"].as<size_t>())
|
|
, m_cOffset(args["offset-c"].as<size_t>())
|
|
, m_dOffset(args["offset-d"].as<size_t>())
|
|
, m_f32XdlMathOp(DataType::Float)
|
|
, m_stochasticRounding(args["stochastic-rounding"].as<bool>())
|
|
|
|
{
|
|
if(args.count("problem-identifier"))
|
|
ContractionProblem::IdentifierToIndices(
|
|
args["problem-identifier"].as<std::string>(),
|
|
m_freeIndices,
|
|
m_batchIndices,
|
|
m_boundIndices,
|
|
m_aOps,
|
|
m_bOps,
|
|
m_cOps,
|
|
m_dOps);
|
|
|
|
if(args.count("type"))
|
|
{
|
|
m_aType = m_bType = m_cType = m_dType = m_alphaType = m_betaType
|
|
= args["type"].as<DataType>();
|
|
}
|
|
|
|
if(args.count("a-type"))
|
|
m_aType = args["a-type"].as<DataType>();
|
|
if(args.count("b-type"))
|
|
m_bType = args["b-type"].as<DataType>();
|
|
if(args.count("c-type"))
|
|
m_cType = args["c-type"].as<DataType>();
|
|
if(args.count("d-type"))
|
|
m_dType = args["d-type"].as<DataType>();
|
|
if(args.count("alpha-type"))
|
|
m_alphaType = args["alpha-type"].as<DataType>();
|
|
if(args.count("beta-type"))
|
|
m_betaType = args["beta-type"].as<DataType>();
|
|
if(args.count("f32-xdl-math-op"))
|
|
m_f32XdlMathOp = args["f32-xdl-math-op"].as<DataType>();
|
|
|
|
m_beta = DataInitialization::getValue<double>(args["init-beta"].as<InitMode>());
|
|
m_alpha = DataInitialization::getValue<double>(args["init-alpha"].as<InitMode>());
|
|
|
|
if(args["convolution-vs-contraction"].as<bool>())
|
|
m_convProblemSizes
|
|
= args["convolution-problem"].as<std::vector<std::vector<size_t>>>();
|
|
|
|
m_problems = createProblems();
|
|
}
|
|
|
|
ClientProblemFactory::~ClientProblemFactory() = default;
|
|
|
|
std::vector<ContractionProblem> const& ClientProblemFactory::problems() const
|
|
{
|
|
return m_problems;
|
|
}
|
|
|
|
std::vector<ContractionProblem> ClientProblemFactory::createProblems()
|
|
{
|
|
std::vector<ContractionProblem> rv;
|
|
rv.reserve(m_problemSizes.size());
|
|
|
|
std::vector<size_t> aStrides, bStrides, cStrides, dStrides;
|
|
|
|
if(m_aStrides.size() == 1)
|
|
aStrides = m_aStrides[0];
|
|
if(m_bStrides.size() == 1)
|
|
bStrides = m_bStrides[0];
|
|
if(m_cStrides.size() == 1)
|
|
cStrides = m_cStrides[0];
|
|
if(m_dStrides.size() == 1)
|
|
dStrides = m_dStrides[0];
|
|
|
|
for(int i = 0; i < m_problemSizes.size(); i++)
|
|
{
|
|
if(m_aStrides.size() == m_problemSizes.size())
|
|
aStrides = m_aStrides[i];
|
|
if(m_bStrides.size() == m_problemSizes.size())
|
|
bStrides = m_bStrides[i];
|
|
if(m_cStrides.size() == m_problemSizes.size())
|
|
cStrides = m_cStrides[i];
|
|
if(m_dStrides.size() == m_problemSizes.size())
|
|
dStrides = m_dStrides[i];
|
|
|
|
rv.push_back(ContractionProblem::FromIndexSizes(m_freeIndices,
|
|
m_batchIndices,
|
|
m_boundIndices,
|
|
m_problemSizes[i],
|
|
m_aType,
|
|
aStrides,
|
|
m_aOps,
|
|
m_aOffset,
|
|
m_bType,
|
|
bStrides,
|
|
m_bOps,
|
|
m_bOffset,
|
|
m_cType,
|
|
cStrides,
|
|
m_cOps,
|
|
m_cOffset,
|
|
m_dType,
|
|
dStrides,
|
|
m_dOps,
|
|
m_dOffset,
|
|
m_beta));
|
|
|
|
rv.back().setAlphaRestriction(toScalarValueEnum(m_alpha));
|
|
rv.back().setCEqualsD(m_cEqualsD);
|
|
|
|
if(i < m_aZeroPads.size())
|
|
{
|
|
const auto& zp = m_aZeroPads[i];
|
|
if(zp.size() % 4 != 0)
|
|
throw std::runtime_error("zero-pad must contain tuples of 4 values");
|
|
for(int zi = 0; zi < zp.size(); zi += 4)
|
|
{
|
|
rv.back().addAZeroPad(
|
|
ContractionProblem::ZeroPad({static_cast<int32_t>(zp[zi + 0]),
|
|
static_cast<int32_t>(zp[zi + 1]),
|
|
static_cast<int64_t>(zp[zi + 2]),
|
|
static_cast<int64_t>(zp[zi + 3])}));
|
|
}
|
|
}
|
|
if(i < m_bZeroPads.size())
|
|
{
|
|
const auto& zp = m_bZeroPads[i];
|
|
if(zp.size() % 4 != 0)
|
|
throw std::runtime_error("zero-pad must contain tuples of 4 values");
|
|
for(int zi = 0; zi < zp.size(); zi += 4)
|
|
{
|
|
rv.back().addBZeroPad(
|
|
ContractionProblem::ZeroPad({static_cast<int32_t>(zp[zi + 0]),
|
|
static_cast<int32_t>(zp[zi + 1]),
|
|
static_cast<int64_t>(zp[zi + 2]),
|
|
static_cast<int64_t>(zp[zi + 3])}));
|
|
}
|
|
}
|
|
rv.back().setAlphaType(m_alphaType);
|
|
rv.back().setBetaType(m_betaType);
|
|
rv.back().setStridedBatched(m_stridedBatched);
|
|
rv.back().setHighPrecisionAccumulate(m_highPrecisionAccumulate);
|
|
rv.back().setKernelLanguage(m_kernelLanguage);
|
|
rv.back().setPerformanceMetric(m_performanceMetric);
|
|
rv.back().setDeterministicMode(m_deterministicMode);
|
|
rv.back().setArithmeticUnit(m_arithmeticUnit);
|
|
rv.back().setFp16AltImpl(m_fp16AltImpl);
|
|
rv.back().setFp16AltImplRound(m_fp16AltImplRound);
|
|
rv.back().setStochasticRounding(m_stochasticRounding);
|
|
rv.back().setF32XdlMathOp(m_f32XdlMathOp);
|
|
|
|
if(m_convProblemSizes.size())
|
|
rv.back().setConvProblemSizes(m_convProblemSizes[i]);
|
|
}
|
|
|
|
return rv;
|
|
}
|
|
} // namespace Client
|
|
} // namespace Tensile
|