squall/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
c2b80ad4e4
cutlass/cuBLAS.cmake

126 lines
2.7 KiB
CMake
Raw Normal View History

CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
message(STATUS "Configuring cublas ...")
CUTLASS 2.1 (#83) CUTLASS 2.1 contributes: - BLAS-style host-side API added to CUTLASS Library - Planar Complex GEMM kernels targeting Volta and Turing Tensor Cores - Minor enhancements and bug fixes
2020-04-08 04:51:25 +08:00
if((DEFINED CUTLASS_ENABLE_CUBLAS AND NOT CUTLASS_ENABLE_CUBLAS) OR
(DEFINED CUBLAS_ENABLED AND NOT CUBLAS_ENABLED))
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
# Don't add cuBLAS if it's defined and false, assume it's not found.
set(CUBLAS_FOUND OFF)
message(STATUS "cuBLAS Disabled.")
elseif(NOT TARGET cublas)
CUTLASS 2.2 (#96) Adds support for NVIDIA Ampere Architecture features. CUDA 11 Toolkit recommended.
2020-06-09 07:17:35 +08:00
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
find_path(
CUTLASS 2.2 (#96) Adds support for NVIDIA Ampere Architecture features. CUDA 11 Toolkit recommended.
2020-06-09 07:17:35 +08:00
_CUBLAS_INCLUDE_DIR
NAMES cublas.h
HINTS
${CUBLAS_INCLUDE_PATH}
ENV CUBLAS_INCLUDE_PATH
${CUBLAS_PATH}
ENV CUBLAS_PATH
${CUDA_TOOLKIT_ROOT_DIR}
PATH_SUFFIXES
include
)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
find_library(
CUTLASS 2.2 (#96) Adds support for NVIDIA Ampere Architecture features. CUDA 11 Toolkit recommended.
2020-06-09 07:17:35 +08:00
_CUBLAS_LIBRARY
NAMES cublas
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
HINTS
CUTLASS 2.2 (#96) Adds support for NVIDIA Ampere Architecture features. CUDA 11 Toolkit recommended.
2020-06-09 07:17:35 +08:00
${CUBLAS_LIBRARY_PATH}
ENV CUBLAS_LIBRARY_PATH
${_CUBLAS_INCLUDE_DIR}/..
${CUBLAS_PATH}
ENV CUBLAS_PATH
${CUDA_TOOLKIT_ROOT_DIR}
PATH_SUFFIXES
lib64
lib/x64
lib
)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
if(_CUBLAS_INCLUDE_DIR AND _CUBLAS_LIBRARY)
message(STATUS "cuBLAS: ${_CUBLAS_LIBRARY}")
message(STATUS "cuBLAS: ${_CUBLAS_INCLUDE_DIR}")
set(CUBLAS_FOUND ON CACHE INTERNAL "cublas Library Found")
set(CUBLAS_LIBRARY ${_CUBLAS_LIBRARY})
set(CUBLAS_INCLUDE_DIR ${_CUBLAS_INCLUDE_DIR})
else()
message(STATUS "cublas not found.")
set(CUBLAS_FOUND OFF CACHE INTERNAL "cublas Library Found")
endif()
endif()
set(CUTLASS_ENABLE_CUBLAS ${CUBLAS_FOUND} CACHE BOOL "Enable CUTLASS to build with cuBLAS library.")
if(CUTLASS_ENABLE_CUBLAS AND NOT CUBLAS_FOUND)
message(FATAL_ERROR "CUTLASS_ENABLE_CUBLAS enabled but cuBLAS library could not be found.")
endif()
if(CUTLASS_ENABLE_CUBLAS AND NOT TARGET cublas)
if(WIN32)
CUTLASS 2.1 (#83) CUTLASS 2.1 contributes: - BLAS-style host-side API added to CUTLASS Library - Planar Complex GEMM kernels targeting Volta and Turing Tensor Cores - Minor enhancements and bug fixes
2020-04-08 04:51:25 +08:00
add_library(cublas STATIC IMPORTED GLOBAL)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
else()
CUTLASS 2.1 (#83) CUTLASS 2.1 contributes: - BLAS-style host-side API added to CUTLASS Library - Planar Complex GEMM kernels targeting Volta and Turing Tensor Cores - Minor enhancements and bug fixes
2020-04-08 04:51:25 +08:00
add_library(cublas SHARED IMPORTED GLOBAL)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
endif()
CUTLASS 2.1 (#83) CUTLASS 2.1 contributes: - BLAS-style host-side API added to CUTLASS Library - Planar Complex GEMM kernels targeting Volta and Turing Tensor Cores - Minor enhancements and bug fixes
2020-04-08 04:51:25 +08:00
add_library(nvidia::cublas ALIAS cublas)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
set_property(
TARGET cublas
PROPERTY IMPORTED_LOCATION
${CUBLAS_LIBRARY})
target_include_directories(
cublas
INTERFACE
$<INSTALL_INTERFACE:include>
$<BUILD_INTERFACE:${CUBLAS_INCLUDE_DIR}>)
find_library(
CUTLASS 2.2 (#96) Adds support for NVIDIA Ampere Architecture features. CUDA 11 Toolkit recommended.
2020-06-09 07:17:35 +08:00
_CUBLASLT_LIBRARY
NAMES cublasLt
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
HINTS
CUTLASS 2.2 (#96) Adds support for NVIDIA Ampere Architecture features. CUDA 11 Toolkit recommended.
2020-06-09 07:17:35 +08:00
${CUBLAS_LIBRARY_PATH}
ENV CUBLAS_LIBRARY_PATH
${_CUBLAS_INCLUDE_DIR}/..
${CUBLAS_PATH}
ENV CUBLAS_PATH
${CUDA_TOOLKIT_ROOT_DIR}
PATH_SUFFIXES
lib64
lib/x64
lib
)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
CUTLASS 2.1 (#83) CUTLASS 2.1 contributes: - BLAS-style host-side API added to CUTLASS Library - Planar Complex GEMM kernels targeting Volta and Turing Tensor Cores - Minor enhancements and bug fixes
2020-04-08 04:51:25 +08:00
if(_CUBLASLT_LIBRARY AND NOT TARGET cublasLt)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
if(WIN32)
CUTLASS 2.1 (#83) CUTLASS 2.1 contributes: - BLAS-style host-side API added to CUTLASS Library - Planar Complex GEMM kernels targeting Volta and Turing Tensor Cores - Minor enhancements and bug fixes
2020-04-08 04:51:25 +08:00
add_library(cublasLt STATIC IMPORTED GLOBAL)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
else()
CUTLASS 2.1 (#83) CUTLASS 2.1 contributes: - BLAS-style host-side API added to CUTLASS Library - Planar Complex GEMM kernels targeting Volta and Turing Tensor Cores - Minor enhancements and bug fixes
2020-04-08 04:51:25 +08:00
add_library(cublasLt SHARED IMPORTED GLOBAL)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
endif()
set_property(
TARGET cublasLt
PROPERTY IMPORTED_LOCATION
${_CUBLASLT_LIBRARY})
CUTLASS 2.1 (#83) CUTLASS 2.1 contributes: - BLAS-style host-side API added to CUTLASS Library - Planar Complex GEMM kernels targeting Volta and Turing Tensor Cores - Minor enhancements and bug fixes
2020-04-08 04:51:25 +08:00
add_library(nvidia::cublasLt ALIAS cublasLt)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
CUTLASS 2.2 (#96) Adds support for NVIDIA Ampere Architecture features. CUDA 11 Toolkit recommended.
2020-06-09 07:17:35 +08:00
target_link_libraries(cublas INTERFACE cublasLt)
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
endif()
endif()
message(STATUS "Configuring cuBLAS ... done.")
Reference in New Issue Copy Permalink