877bdcace6
CUTLASS 1.3 Release - Efficient GEMM kernel targeting Volta Tensor Cores via mma.sync instruction added in CUDA 10.1.
155 lines
4.8 KiB
CMake
155 lines
4.8 KiB
CMake
# Copyright (c) 2017-2019, NVIDIA CORPORATION. All rights reserved.
|
|
#
|
|
# Redistribution and use in source and binary forms, with or without modification, are permitted
|
|
# provided that the following conditions are met:
|
|
# * Redistributions of source code must retain the above copyright notice, this list of
|
|
# conditions and the following disclaimer.
|
|
# * Redistributions in binary form must reproduce the above copyright notice, this list of
|
|
# conditions and the following disclaimer in the documentation and/or other materials
|
|
# provided with the distribution.
|
|
# * Neither the name of the NVIDIA CORPORATION nor the names of its contributors may be used
|
|
# to endorse or promote products derived from this software without specific prior written
|
|
# permission.
|
|
#
|
|
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR
|
|
# IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
|
|
# FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NVIDIA CORPORATION BE LIABLE
|
|
# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
|
|
# BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
|
|
# OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
|
|
# STRICT LIABILITY, OR TOR (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
|
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
|
|
|
|
include_directories(
|
|
../../external/googletest/googletest
|
|
.
|
|
)
|
|
|
|
link_directories(
|
|
../../external/googletest/googletest
|
|
)
|
|
|
|
link_libraries(
|
|
gtest
|
|
)
|
|
|
|
set(CUTLASS_UNIT_TEST_HEADERS
|
|
cutlass_unit_test.h
|
|
core/layout_verification.h
|
|
gemm/run_gemm.h
|
|
gemm/gemm_testbed.h
|
|
reduction/batched_reduction_testbed.h
|
|
reduction/test_batched_reduction.h
|
|
)
|
|
|
|
set(CUTLASS_UNIT_TEST_SOURCES
|
|
cutlass_unit_test.cpp
|
|
tile_iterator_test.cu
|
|
core/tensor_ref.cu
|
|
core/tensor_view.cu
|
|
util/unique_ptr.cu
|
|
util/host_tensor.cu
|
|
util/complex.cu
|
|
util/tensor_foreach.cu
|
|
util/tensor_elementwise.cu
|
|
core/layout_verification.cu
|
|
core/predicate_vector.cu
|
|
core/tile_iterator.cu
|
|
core/zip_tile_iterator.cu
|
|
gemm/dgemm.cu
|
|
gemm/hgemm_128x128x8.cu
|
|
gemm/hgemm_128x128x16.cu
|
|
gemm/hgemm_128x32x8.cu
|
|
gemm/hgemm_128x64x8.cu
|
|
gemm/igemm_128x128x32.cu
|
|
gemm/igemm_128x64x32.cu
|
|
gemm/igemm_128x32x32.cu
|
|
gemm/igemm_128x128x32_float.cu
|
|
gemm/igemm_128x128x32_int8.cu
|
|
gemm/igemm_32x32x128.cu
|
|
gemm/sgemm_128x128x8.cu
|
|
gemm/sgemm_128x128x16.cu
|
|
gemm/sgemm_128x64x8.cu
|
|
gemm/sgemm_128x64x16.cu
|
|
gemm/sgemm_128x32x8.cu
|
|
gemm/sgemm_128x32x16.cu
|
|
gemm/sgemm_64x128x8.cu
|
|
gemm/sgemm_64x128x16.cu
|
|
gemm/sgemm_64x64x8.cu
|
|
gemm/sgemm_64x64x16.cu
|
|
gemm/sgemm_64x32x8.cu
|
|
gemm/sgemm_64x32x16.cu
|
|
gemm/sgemm_32x128x8.cu
|
|
gemm/fp16_sgemm_fp32_128x128x16.cu
|
|
gemm/fp16_sgemm_fp16_128x128x16.cu
|
|
gemm/wmma_gemm.cu
|
|
gemm/fp16_wmma_gemm_fp16.cu
|
|
gemm/wmma_gemm_non_multiple16.cu
|
|
gemm/fp16_wmma_gemm_fp16_non_multiple16.cu
|
|
gemm/wmma_binary_gemm.cu
|
|
gemm/wmma_integer_gemm.cu
|
|
gemm/sgemm_threadblock_swizzle_nn.cu
|
|
gemm/sgemm_threadblock_swizzle_nt.cu
|
|
gemm/sgemm_threadblock_swizzle_tn.cu
|
|
gemm/sgemm_threadblock_swizzle_tt.cu
|
|
gemm/batched_strided_sgemm_128x128x8.cu
|
|
gemm/batched_strided_dgemm_128x128x8.cu
|
|
gemm/batched_strided_hgemm_128x128x8.cu
|
|
gemm/batched_strided_wmma_gemm.cu
|
|
gemm/batched_strided_fp16_wmma_gemm_fp16.cu
|
|
gemm/batched_strided_volta884_hgemm.cu
|
|
gemm/epilogue_functor.cu
|
|
gemm/volta884_gemm_epilogue.cu
|
|
gemm/volta884_h884gemm.cu
|
|
gemm/volta884_gemm.cu
|
|
gemm/volta884_gemm_threadblock_swizzle.cu
|
|
gemm/volta884_h884gemm_threadblock_swizzle.cu
|
|
reduction/batched_reduction.cu
|
|
reduction/mixed_batched_reduction.cu
|
|
gemm/splitK_sgemm.cu
|
|
gemm/splitK_igemm.cu
|
|
gemm/splitK_fp16_sgemm_fp16.cu
|
|
gemm/splitK_dgemm.cu
|
|
gemm/splitK_hgemm.cu
|
|
gemm/splitK_wmma_gemm.cu
|
|
gemm/splitK_volta884_hgemm.cu
|
|
gemm/partitionedK_sgemm_128x128x8.cu
|
|
gemm/partitionedK_volta884_hgemm.cu
|
|
)
|
|
|
|
if (CUTLASS_NVRTC_ENABLE)
|
|
list(APPEND CUTLASS_UNIT_TEST_SOURCES gemm/gemm_nvrtc.cu)
|
|
endif()
|
|
|
|
source_group("Source\ Files" FILES ${CUTLASS_UNIT_TEST_SOURCES})
|
|
|
|
if (NOT CUTLASS_NATIVE_CUDA)
|
|
# cuda_add_executable does not take interface include directories into account
|
|
# Let's fetch them and pass them to CUDA.
|
|
get_target_property(CUTLASS_INCLUDES CUTLASS INTERFACE_INCLUDE_DIRECTORIES)
|
|
include_directories("${CUTLASS_INCLUDES}")
|
|
endif()
|
|
|
|
cutlass_add_executable(
|
|
cutlass_unit_test
|
|
${CUTLASS_UNIT_TEST_SOURCES}
|
|
${CUTLASS_UNIT_TEST_HEADERS}
|
|
)
|
|
|
|
if (CUTLASS_NVRTC_ENABLE)
|
|
target_link_libraries(cutlass_unit_test
|
|
cutlass_nvrtc
|
|
cuda
|
|
)
|
|
|
|
if (WIN32)
|
|
target_link_libraries(cutlass_unit_test Dbghelp)
|
|
endif()
|
|
endif()
|
|
|
|
if(CUTLASS_ENABLE_CUBLAS)
|
|
target_link_libraries(cutlass_unit_test ${CUBLAS_LIBRARY})
|
|
endif()
|
|
|