Merge commit 'f6c2ff9dcedbc58065ae1fc10a661f00716c6839' into develop

This commit is contained in:
assistant-librarian[bot]
2025-06-16 14:09:09 +00:00
parent 4501568487
commit 6fc36b26e6
41 changed files with 2103 additions and 106 deletions

View File

@@ -252,7 +252,7 @@ add_subdirectory(reduce)
add_subdirectory(convnd_fwd)
add_subdirectory(convnd_bwd_data)
add_subdirectory(grouped_convnd_fwd)
add_subdirectory(grouped_convnd_fwd_bias_clamp)
add_subdirectory(grouped_convnd_fwd_activation)
add_subdirectory(grouped_convnd_bwd_weight)
add_subdirectory(block_to_ctile_map)
add_subdirectory(softmax)

View File

@@ -0,0 +1,10 @@
if(GPU_TARGETS MATCHES "gfx9")
add_gtest_executable(test_grouped_convnd_fwd_bias_clamp test_grouped_convnd_fwd_bias_clamp.cpp)
target_link_libraries(test_grouped_convnd_fwd_bias_clamp PRIVATE utility device_grouped_conv2d_fwd_bias_clamp_instance device_grouped_conv3d_fwd_bias_clamp_instance)
add_gtest_executable(test_grouped_convnd_fwd_gk_bias_clamp test_grouped_convnd_fwd_gk_bias_clamp.cpp)
target_link_libraries(test_grouped_convnd_fwd_gk_bias_clamp PRIVATE utility device_grouped_conv2d_fwd_bias_clamp_instance device_grouped_conv3d_fwd_bias_clamp_instance)
add_gtest_executable(test_grouped_convnd_fwd_clamp test_grouped_convnd_fwd_clamp.cpp)
target_link_libraries(test_grouped_convnd_fwd_clamp PRIVATE utility device_grouped_conv2d_fwd_clamp_instance device_grouped_conv3d_fwd_clamp_instance)
endif()

View File

@@ -41,7 +41,8 @@ class TestGroupedConvndFwd : public ::testing::Test
DataType,
DataType,
DataType,
IndexType>(
IndexType,
false /*BiasGK*/>(
true, // do_verification
1, // init_method: integer value
false, // do_log

View File

@@ -0,0 +1,95 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.
#include <cstdlib>
#include <iostream>
#include <initializer_list>
#include <vector>
#include <gtest/gtest.h>
#include "profiler/profile_grouped_conv_fwd_impl.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
using Clamp = ck::tensor_operation::element_wise::Clamp;
template <typename Tuple>
class TestGroupedConvndFwd : public ::testing::Test
{
protected:
using DataType = std::tuple_element_t<0, Tuple>;
using InLayout = std::tuple_element_t<1, Tuple>;
using WeiLayout = std::tuple_element_t<2, Tuple>;
using OutLayout = std::tuple_element_t<3, Tuple>;
using IndexType = ck::index_t;
std::vector<ck::utils::conv::ConvParam> conv_params;
template <ck::index_t NDimSpatial>
void Run()
{
EXPECT_FALSE(conv_params.empty());
bool pass = true;
Clamp out_element_op{0.f, 256.f};
for(auto& param : conv_params)
{
pass = pass && ck::profiler::profile_grouped_conv_fwd_impl<NDimSpatial,
InLayout,
WeiLayout,
OutLayout,
DataType,
DataType,
DataType,
DataType,
DataType,
IndexType,
Clamp>(
true, // do_verification
1, // init_method: integer value
false, // do_log
false, // time_kernel
param,
out_element_op);
}
EXPECT_TRUE(pass);
}
};
using namespace ck::tensor_layout::convolution;
using KernelTypes2d = ::testing::Types<std::tuple<ck::bhalf_t, NHWGC, GKYXC, NHWGK>>;
using KernelTypes3d = ::testing::Types<std::tuple<ck::bhalf_t, NDHWGC, GKZYXC, NDHWGK>>;
template <typename Tuple>
class TestGroupedConvndFwd2d : public TestGroupedConvndFwd<Tuple>
{
};
template <typename Tuple>
class TestGroupedConvndFwd3d : public TestGroupedConvndFwd<Tuple>
{
};
TYPED_TEST_SUITE(TestGroupedConvndFwd2d, KernelTypes2d);
TYPED_TEST_SUITE(TestGroupedConvndFwd3d, KernelTypes3d);
TYPED_TEST(TestGroupedConvndFwd2d, Test2D)
{
this->conv_params.clear();
this->conv_params.push_back(
{2, 2, 32, 128, 256, {1, 1}, {7, 7}, {2, 2}, {1, 1}, {0, 0}, {0, 0}});
this->conv_params.push_back(
{2, 2, 32, 128, 256, {3, 3}, {14, 14}, {1, 1}, {1, 1}, {1, 1}, {1, 1}});
this->template Run<2>();
}
TYPED_TEST(TestGroupedConvndFwd3d, Test3D)
{
this->conv_params.clear();
this->conv_params.push_back(
{3, 2, 32, 128, 256, {1, 1, 1}, {7, 7, 7}, {2, 2, 2}, {1, 1, 1}, {0, 0, 0}, {0, 0, 0}});
this->conv_params.push_back(
{3, 2, 32, 128, 256, {3, 3, 3}, {14, 14, 3}, {1, 1, 1}, {1, 1, 1}, {1, 1, 1}, {1, 1, 1}});
this->template Run<3>();
}

View File

@@ -0,0 +1,93 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.
#include <cstdlib>
#include <iostream>
#include <initializer_list>
#include <vector>
#include <gtest/gtest.h>
#include "profiler/profile_grouped_conv_fwd_bias_clamp_impl.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
using AddClamp = ck::tensor_operation::element_wise::AddClamp;
template <typename Tuple>
class TestGroupedConvndFwd : public ::testing::Test
{
protected:
using DataType = std::tuple_element_t<0, Tuple>;
using InLayout = std::tuple_element_t<1, Tuple>;
using WeiLayout = std::tuple_element_t<2, Tuple>;
using OutLayout = std::tuple_element_t<3, Tuple>;
using IndexType = ck::index_t;
std::vector<ck::utils::conv::ConvParam> conv_params;
template <ck::index_t NDimSpatial>
void Run()
{
EXPECT_FALSE(conv_params.empty());
bool pass = true;
for(auto& param : conv_params)
{
pass = pass && ck::profiler::profile_grouped_conv_fwd_bias_clamp_impl<NDimSpatial,
InLayout,
WeiLayout,
OutLayout,
DataType,
DataType,
DataType,
DataType,
DataType,
IndexType,
true /*BiasGK*/>(
true, // do_verification
1, // init_method: integer value
false, // do_log
false, // time_kernel
param);
}
EXPECT_TRUE(pass);
}
};
using namespace ck::tensor_layout::convolution;
using KernelTypes2d = ::testing::Types<std::tuple<ck::bhalf_t, NHWGC, GKYXC, NHWGK>>;
using KernelTypes3d = ::testing::Types<std::tuple<ck::bhalf_t, NDHWGC, GKZYXC, NDHWGK>>;
template <typename Tuple>
class TestGroupedConvndFwd2d : public TestGroupedConvndFwd<Tuple>
{
};
template <typename Tuple>
class TestGroupedConvndFwd3d : public TestGroupedConvndFwd<Tuple>
{
};
TYPED_TEST_SUITE(TestGroupedConvndFwd2d, KernelTypes2d);
TYPED_TEST_SUITE(TestGroupedConvndFwd3d, KernelTypes3d);
TYPED_TEST(TestGroupedConvndFwd2d, Test2D)
{
this->conv_params.clear();
this->conv_params.push_back(
{2, 2, 32, 128, 256, {1, 1}, {7, 7}, {2, 2}, {1, 1}, {0, 0}, {0, 0}});
this->conv_params.push_back(
{2, 2, 32, 128, 256, {3, 3}, {14, 14}, {1, 1}, {1, 1}, {1, 1}, {1, 1}});
this->template Run<2>();
}
TYPED_TEST(TestGroupedConvndFwd3d, Test3D)
{
this->conv_params.clear();
this->conv_params.push_back(
{3, 2, 32, 128, 256, {1, 1, 1}, {7, 7, 7}, {2, 2, 2}, {1, 1, 1}, {0, 0, 0}, {0, 0, 0}});
this->conv_params.push_back(
{3, 2, 32, 128, 256, {3, 3, 3}, {14, 14, 3}, {1, 1, 1}, {1, 1, 1}, {1, 1, 1}, {1, 1, 1}});
this->template Run<3>();
}

View File

@@ -1,4 +0,0 @@
if(GPU_TARGETS MATCHES "gfx9")
add_gtest_executable(test_grouped_convnd_fwd_bias_clamp test_grouped_convnd_fwd_bias_clamp.cpp)
target_link_libraries(test_grouped_convnd_fwd_bias_clamp PRIVATE utility device_grouped_conv2d_fwd_bias_clamp_instance device_grouped_conv3d_fwd_bias_clamp_instance)
endif()