CMakeLists.txt 23.2 KB
Newer Older
1
cmake_minimum_required(VERSION 3.14)
2
3
4
5
if(POLICY CMP0140)
  # policies CMP0140 not known to CMake until 3.25
  cmake_policy(SET CMP0140 NEW)
endif()
JD's avatar
JD committed
6

7
8
get_property(_GENERATOR_IS_MULTI_CONFIG GLOBAL PROPERTY GENERATOR_IS_MULTI_CONFIG)

9
10
# This has to be initialized before the project() command appears
# Set the default of CMAKE_BUILD_TYPE to be release, unless user specifies with -D.  MSVC_IDE does not use CMAKE_BUILD_TYPE
11
12
13
14
15
16
if(_GENERATOR_IS_MULTI_CONFIG)
    set(CMAKE_CONFIGURATION_TYPES "Debug;Release;RelWithDebInfo;MinSizeRel" CACHE STRING
            "Available build types (configurations) on multi-config generators")
else()
    set(CMAKE_BUILD_TYPE Release CACHE STRING
            "Choose the type of build, options are: None Debug Release RelWithDebInfo MinSizeRel.")
17
18
19
endif()

# Default installation path
20
if(NOT WIN32)
21
22
23
    set(CMAKE_INSTALL_PREFIX "/opt/rocm" CACHE PATH "")
endif()

24
set(version 1.1.0)
JD's avatar
JD committed
25
# Check support for CUDA/HIP in Cmake
26
project(composable_kernel VERSION ${version} LANGUAGES CXX HIP)
27
include(CTest)
Chao Liu's avatar
Chao Liu committed
28

29
30
31
# Usage: for customized Python location cmake -DCK_USE_ALTERNATIVE_PYTHON="/opt/Python-3.8.13/bin/python3.8"
# CK Codegen requires dataclass which is added in Python 3.7
# Python version 3.8 is required for general good practice as it is default for Ubuntu 20.04
32
if(NOT CK_USE_ALTERNATIVE_PYTHON)
33
   find_package(Python3 3.8 COMPONENTS Interpreter REQUIRED)
34
35
36
else()
   message("Using alternative python version")
   set(EXTRA_PYTHON_PATH)
37
   # this is overly restrictive, we may need to be more flexible on the following
38
39
40
41
42
43
44
45
   string(REPLACE "/bin/python3.8" "" EXTRA_PYTHON_PATH "${CK_USE_ALTERNATIVE_PYTHON}")
   message("alternative python path is: ${EXTRA_PYTHON_PATH}")
   find_package(Python3 3.6 COMPONENTS Interpreter REQUIRED)
   add_definitions(-DPython3_EXECUTABLE="${CK_USE_ALTERNATIVE_PYTHON}")
   set(Python3_EXECUTABLE "${CK_USE_ALTERNATIVE_PYTHON}")
   set(PYTHON_EXECUTABLE "${CK_USE_ALTERNATIVE_PYTHON}")
   set(ENV{LD_LIBRARY_PATH} "${EXTRA_PYTHON_PATH}/lib:$ENV{LD_LIBRARY_PATH}")
endif()
carlushuang's avatar
carlushuang committed
46

47
48
list(APPEND CMAKE_MODULE_PATH "${PROJECT_SOURCE_DIR}/cmake")

49
if (DTYPES)
50
51
52
53
54
55
56
57
    add_definitions(-DDTYPES)
    if (DTYPES MATCHES "int8")
        add_definitions(-DCK_ENABLE_INT8)
        set(CK_ENABLE_INT8 "ON")
    endif()
    if (DTYPES MATCHES "fp8")
        add_definitions(-DCK_ENABLE_FP8)
        set(CK_ENABLE_FP8 "ON")
58
59
60
61
    endif()
    if (DTYPES MATCHES "bf8")
        add_definitions(-DCK_ENABLE_BF8)
        set(CK_ENABLE_BF8 "ON")
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
    endif()
    if (DTYPES MATCHES "fp16")
        add_definitions(-DCK_ENABLE_FP16)
        set(CK_ENABLE_FP16 "ON")
    endif()
    if (DTYPES MATCHES "fp32")
        add_definitions(-DCK_ENABLE_FP32)
        set(CK_ENABLE_FP32 "ON")
    endif()
    if (DTYPES MATCHES "fp64")
        add_definitions(-DCK_ENABLE_FP64)
        set(CK_ENABLE_FP64 "ON")
    endif()
    if (DTYPES MATCHES "bf16")
        add_definitions(-DCK_ENABLE_BF16)
        set(CK_ENABLE_BF16 "ON")
    endif()
    message("DTYPES macro set to ${DTYPES}")
80
else()
81
    add_definitions(-DCK_ENABLE_INT8 -DCK_ENABLE_FP16 -DCK_ENABLE_FP32 -DCK_ENABLE_FP64 -DCK_ENABLE_BF16 -DCK_ENABLE_FP8 -DCK_ENABLE_BF8)
82
83
84
85
86
    set(CK_ENABLE_INT8 "ON")
    set(CK_ENABLE_FP16 "ON")
    set(CK_ENABLE_FP32 "ON")
    set(CK_ENABLE_FP64 "ON")
    set(CK_ENABLE_BF16 "ON")
87
88
    set(CK_ENABLE_FP8 "ON")
    set(CK_ENABLE_BF8 "ON")
89
90
endif()

91
92
#for f8/bf8_t type
add_compile_options(-Wno-bit-int-extension)
93
add_compile_options(-Wno-pass-failed)
94
add_compile_options(-Wno-switch-default)
95

96
97
if(DL_KERNELS)
    add_definitions(-DDL_KERNELS)
98
    set(CK_ENABLE_DL_KERNELS "ON")
99
endif()
100
101
102
option(CK_USE_CODEGEN "Enable codegen library" OFF)
if(CK_USE_CODEGEN)
    add_definitions(-DCK_USE_CODEGEN)
103
104
endif()

105
106
include(getopt)

107
108
109
# CK version file to record release version as well as git commit hash
find_package(Git REQUIRED)
execute_process(COMMAND "${GIT_EXECUTABLE}" rev-parse HEAD OUTPUT_VARIABLE COMMIT_ID OUTPUT_STRIP_TRAILING_WHITESPACE)
110
configure_file(include/ck/version.h.in ${CMAKE_CURRENT_BINARY_DIR}/include/ck/version.h)
JD's avatar
JD committed
111

112
set(ROCM_SYMLINK_LIBS OFF)
Anthony Chang's avatar
Anthony Chang committed
113
find_package(ROCM REQUIRED PATHS /opt/rocm)
JD's avatar
JD committed
114
115
116
117
118
119

include(ROCMInstallTargets)
include(ROCMPackageConfigHelpers)
include(ROCMSetupVersion)
include(ROCMInstallSymlinks)
include(ROCMCreatePackage)
Chao Liu's avatar
Chao Liu committed
120
include(CheckCXXCompilerFlag)
121
include(ROCMCheckTargetIds)
JD's avatar
JD committed
122
include(TargetFlags)
123
124
125

rocm_setup_version(VERSION ${version})

126
list(APPEND CMAKE_PREFIX_PATH ${CMAKE_INSTALL_PREFIX} ${CMAKE_INSTALL_PREFIX}/llvm ${CMAKE_INSTALL_PREFIX}/hip /opt/rocm /opt/rocm/llvm /opt/rocm/hip "$ENV{ROCM_PATH}" "$ENV{HIP_PATH}")
JD's avatar
JD committed
127

128
message("GPU_TARGETS= ${GPU_TARGETS}")
129
130
131
132
133
134
message("GPU_ARCHS= ${GPU_ARCHS}")
if(GPU_ARCHS)
    #disable GPU_TARGETS to avoid conflicts, this needs to happen before we call hip package
    unset(GPU_TARGETS CACHE)
    unset(AMDGPU_TARGETS CACHE)
endif()
135
136
137
138
139
if(GPU_TARGETS)
    set(USER_GPU_TARGETS 1)
else()
    set(USER_GPU_TARGETS 0)
endif()
Illia Silin's avatar
Illia Silin committed
140
find_package(hip REQUIRED)
141
142
143
144
145
# No assumption that HIP kernels are launched with uniform block size for backward compatibility
# SWDEV-413293 and https://reviews.llvm.org/D155213
math(EXPR hip_VERSION_FLAT "(${hip_VERSION_MAJOR} * 1000 + ${hip_VERSION_MINOR}) * 100000 + ${hip_VERSION_PATCH}")
message("hip_version_flat=${hip_VERSION_FLAT}")

146
message("checking which targets are supported")
147
#In order to build just the CK library (without tests and examples) for all supported GPU targets
148
#use -D GPU_ARCHS="gfx908;gfx90a;gfx942;gfx1030;gfx1100;gfx1101;gfx1102;gfx1200;gfx1201"
149
150
151
152
153
154
#the GPU_TARGETS flag will be reset in this case in order to avoid conflicts.
#
#In order to build CK along with all tests and examples it should be OK to set GPU_TARGETS to just 1 or 2 similar architectures.
if(NOT ENABLE_ASAN_PACKAGING)
    if(NOT WIN32 AND ${hip_VERSION_FLAT} LESS 600300000)
        # WORKAROUND: compiler does not yet fully support gfx12 targets, need to fix version above
illsilin's avatar
illsilin committed
155
        set(CK_GPU_TARGETS "gfx950")
156
    else()
illsilin's avatar
illsilin committed
157
        set(CK_GPU_TARGETS "gfx950")
158
    endif()
159
else()
160
    #build CK only for xnack-supported targets when using ASAN
161
    set(CK_GPU_TARGETS "gfx908:xnack+;gfx90a:xnack+;gfx942:xnack+")
162
163
endif()

164
165
166
167
#if user set GPU_ARCHS on the cmake command line, overwrite default target list with user's list
#otherwise, if user set GPU_TARGETS, use that set of targets
if(GPU_ARCHS)
    set(CK_GPU_TARGETS ${GPU_ARCHS})
168
else()
169
    if(USER_GPU_TARGETS)
170
        set(CK_GPU_TARGETS ${GPU_TARGETS})
171
172
    endif()
endif()
Illia Silin's avatar
Illia Silin committed
173
174
175
176
#if the user did not set GPU_TARGETS, delete whatever was set by HIP package
if(NOT USER_GPU_TARGETS)
    set(GPU_TARGETS "")
endif()
177
178
179
#make sure all the targets on the list are actually supported by the current compiler
rocm_check_target_ids(SUPPORTED_GPU_TARGETS
        TARGETS ${CK_GPU_TARGETS})
180

181
message("Building CK for the following targets: ${SUPPORTED_GPU_TARGETS}")
182

183
if (SUPPORTED_GPU_TARGETS MATCHES "gfx9")
184
185
    message("Enabling XDL instances")
    add_definitions(-DCK_USE_XDL)
186
    set(CK_USE_XDL "ON")
187
endif()
illsilin's avatar
illsilin committed
188
if (SUPPORTED_GPU_TARGETS MATCHES "gfx94")
189
190
    message("Enabling FP8 gemms on native architectures")
    add_definitions(-DCK_USE_GFX94)
191
    set(CK_USE_GFX94 "ON")
192
endif()
193
194
195
if (SUPPORTED_GPU_TARGETS MATCHES "gfx95")
    add_definitions(-DCK_USE_AMD_MFMA_GFX950)
endif()
196
if (SUPPORTED_GPU_TARGETS MATCHES "gfx11" OR SUPPORTED_GPU_TARGETS MATCHES "gfx12")
197
198
    message("Enabling WMMA instances")
    add_definitions(-DCK_USE_WMMA)
199
    set(CK_USE_WMMA "ON")
200
endif()
illsilin's avatar
illsilin committed
201
if (SUPPORTED_GPU_TARGETS MATCHES "gfx12" OR SUPPORTED_GPU_TARGETS MATCHES "gfx950")
202
203
204
205
206
207
    add_definitions(-DCK_USE_OCP_FP8)
    set(CK_USE_OCP_FP8 "ON")
endif()
if (SUPPORTED_GPU_TARGETS MATCHES "gfx90a" OR SUPPORTED_GPU_TARGETS MATCHES "gfx94")
    add_definitions(-DCK_USE_FNUZ_FP8)
    set(CK_USE_FNUZ_FP8 "ON")
208
209
endif()

Illia Silin's avatar
Illia Silin committed
210
211
212
option(CK_USE_FP8_ON_UNSUPPORTED_ARCH "Enable FP8 GEMM instances on older architectures" OFF)
if(CK_USE_FP8_ON_UNSUPPORTED_ARCH AND (SUPPORTED_GPU_TARGETS MATCHES "gfx90a" OR SUPPORTED_GPU_TARGETS MATCHES "gfx908"))
    add_definitions(-DCK_USE_FP8_ON_UNSUPPORTED_ARCH)
213
    set(CK_USE_FP8_ON_UNSUPPORTED_ARCH "ON")
214
215
endif()

216
217
218
# CK config file to record supported datatypes, etc.
configure_file(include/ck/config.h.in ${CMAKE_CURRENT_BINARY_DIR}/include/ck/config.h)

219
if(NOT WIN32 AND ${hip_VERSION_FLAT} GREATER 500723302)
220
221
222
223
224
  check_cxx_compiler_flag("-fno-offload-uniform-block" HAS_NO_OFFLOAD_UNIFORM_BLOCK)
  if(HAS_NO_OFFLOAD_UNIFORM_BLOCK)
    message("Adding the fno-offload-uniform-block compiler flag")
    add_compile_options(-fno-offload-uniform-block)
  endif()
225
endif()
226
227
228
229
230
231
232
if(NOT WIN32 AND ${hip_VERSION_FLAT} GREATER 500500000)
  check_cxx_compiler_flag("-mllvm --lsr-drop-solution=1" HAS_LSR_DROP_SOLUTION)
  if(HAS_LSR_DROP_SOLUTION)
    message("Adding the lsr-drop-solution=1 compiler flag")
    add_compile_options("SHELL: -mllvm --lsr-drop-solution=1")
  endif()
endif()
233
if(NOT WIN32 AND ${hip_VERSION_FLAT} GREATER 600140090)
234
235
236
237
238
  check_cxx_compiler_flag("-mllvm -enable-post-misched=0" HAS_ENABLE_POST_MISCHED)
  if(HAS_ENABLE_POST_MISCHED)
    message("Adding the enable-post-misched=0 compiler flag")
    add_compile_options("SHELL: -mllvm -enable-post-misched=0")
  endif()
239
endif()
240
241
set(check-coerce)
check_cxx_compiler_flag(" -mllvm -amdgpu-coerce-illegal-types=1" check-coerce)
242
if(NOT WIN32 AND check-coerce AND ${hip_VERSION_FLAT} GREATER 600241132)
243
244
245
246
247
248
249
   message("Adding the amdgpu-coerce-illegal-types=1")
   add_compile_options("SHELL: -mllvm -amdgpu-coerce-illegal-types=1")
endif()
if(NOT WIN32 AND ${hip_VERSION_FLAT} GREATER 600241132)
   message("Adding -amdgpu-early-inline-all=true and -amdgpu-function-calls=false")
   add_compile_options("SHELL: -mllvm -amdgpu-early-inline-all=true")
   add_compile_options("SHELL: -mllvm -amdgpu-function-calls=false")
250
endif()
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
#
# Seperate linking jobs from compiling
# Too many concurrent linking jobs can break the build
# Copied from LLVM
set(CK_PARALLEL_LINK_JOBS "" CACHE STRING
  "Define the maximum number of concurrent link jobs (Ninja only).")
if(CMAKE_GENERATOR MATCHES "Ninja")
  if(CK_PARALLEL_LINK_JOBS)
    set_property(GLOBAL APPEND PROPERTY JOB_POOLS link_job_pool=${CK_PARALLEL_LINK_JOBS})
    set(CMAKE_JOB_POOL_LINK link_job_pool)
  endif()
elseif(CK_PARALLEL_LINK_JOBS)
  message(WARNING "Job pooling is only available with Ninja generators.")
endif()
# Similar for compiling
set(CK_PARALLEL_COMPILE_JOBS "" CACHE STRING
  "Define the maximum number of concurrent compile jobs (Ninja only).")
if(CMAKE_GENERATOR MATCHES "Ninja")
  if(CK_PARALLEL_COMPILE_JOBS)
    set_property(GLOBAL APPEND PROPERTY JOB_POOLS compile_job_pool=${CK_PARALLEL_COMPILE_JOBS})
    set(CMAKE_JOB_POOL_COMPILE compile_job_pool)
  endif()
elseif(CK_PARALLEL_COMPILE_JOBS)
  message(WARNING "Job pooling is only available with Ninja generators.")
endif()


278
option(USE_BITINT_EXTENSION_INT4 "Whether to enable clang's BitInt extension to provide int4 data type." OFF)
Illia Silin's avatar
Illia Silin committed
279
option(USE_OPT_GFX11 "Whether to enable LDS cumode and Wavefront32 mode for GFX11 silicons." OFF)
Adam Osewski's avatar
Adam Osewski committed
280
281
282
283
284
285
286

if(USE_BITINT_EXTENSION_INT4)
    add_compile_definitions(CK_EXPERIMENTAL_BIT_INT_EXTENSION_INT4)
    add_compile_options(-Wno-bit-int-extension)
    message("CK compiled with USE_BITINT_EXTENSION_INT4 set to ${USE_BITINT_EXTENSION_INT4}")
endif()

Illia Silin's avatar
Illia Silin committed
287
if(USE_OPT_GFX11)
288
289
    add_compile_options(-mcumode)
    add_compile_options(-mno-wavefrontsize64)
Illia Silin's avatar
Illia Silin committed
290
    message("CK compiled with USE_OPT_GFX11 set to ${USE_OPT_GFX11}")
291
292
endif()

293
294
295
296
297
## Threads
set(THREADS_PREFER_PTHREAD_FLAG ON)
find_package(Threads REQUIRED)
link_libraries(Threads::Threads)

298
## C++
Chao Liu's avatar
Chao Liu committed
299
set(CMAKE_CXX_STANDARD 17)
Chao Liu's avatar
Chao Liu committed
300
set(CMAKE_CXX_STANDARD_REQUIRED ON)
Chao Liu's avatar
Chao Liu committed
301
set(CMAKE_CXX_EXTENSIONS OFF)
302
303
message("CMAKE_CXX_COMPILER: ${CMAKE_CXX_COMPILER}")

304
305
306
307
308
309
310
311
312
313
# https://gcc.gnu.org/onlinedocs/libstdc++/manual/using_macros.html
# _GLIBCXX_ASSERTIONS
# Undefined by default. When defined, enables extra error checking in the form of
# precondition assertions, such as bounds checking in strings and null pointer
# checks when dereferencing smart pointers
option(USE_GLIBCXX_ASSERTIONS "Turn on additional c++ library checks." OFF)
if(USE_GLIBCXX_ASSERTIONS)
  add_compile_options(-Wp,-D_GLIBCXX_ASSERTIONS)
endif()

314
315
316
317
318
## HIP
set(CMAKE_HIP_PLATFORM amd)
set(CMAKE_HIP_COMPILER ${CMAKE_CXX_COMPILER})
set(CMAKE_HIP_EXTENSIONS ON)
message("CMAKE_HIP_COMPILER: ${CMAKE_HIP_COMPILER}")
Chao Liu's avatar
Chao Liu committed
319

320
## OpenMP
Chao Liu's avatar
Chao Liu committed
321
322
323
324
325
326
327
328
329
330
331
if(CMAKE_CXX_COMPILER_ID MATCHES "Clang")
	# workaround issue hipcc in rocm3.5 cannot find openmp
	set(OpenMP_CXX "${CMAKE_CXX_COMPILER}")
	set(OpenMP_CXX_FLAGS "-fopenmp=libomp -Wno-unused-command-line-argument")
	set(OpenMP_CXX_LIB_NAMES "libomp" "libgomp" "libiomp5")
	set(OpenMP_libomp_LIBRARY ${OpenMP_CXX_LIB_NAMES})
	set(OpenMP_libgomp_LIBRARY ${OpenMP_CXX_LIB_NAMES})
	set(OpenMP_libiomp5_LIBRARY ${OpenMP_CXX_LIB_NAMES})
else()
	find_package(OpenMP REQUIRED)
endif()
Chao Liu's avatar
Chao Liu committed
332

Chao Liu's avatar
Chao Liu committed
333
334
335
336
message("OpenMP_CXX_LIB_NAMES: ${OpenMP_CXX_LIB_NAMES}")
message("OpenMP_gomp_LIBRARY: ${OpenMP_gomp_LIBRARY}")
message("OpenMP_pthread_LIBRARY: ${OpenMP_pthread_LIBRARY}")
message("OpenMP_CXX_FLAGS: ${OpenMP_CXX_FLAGS}")
Chao Liu's avatar
Chao Liu committed
337

Chao Liu's avatar
Chao Liu committed
338
339
link_libraries(${OpenMP_gomp_LIBRARY})
link_libraries(${OpenMP_pthread_LIBRARY})
Chao Liu's avatar
Chao Liu committed
340

341
## HIP
JD's avatar
JD committed
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
# Override HIP version in config.h, if necessary.
# The variables set by find_package() can't be overwritten,
# therefore let's use intermediate variables.
set(CK_HIP_VERSION_MAJOR "${HIP_VERSION_MAJOR}")
set(CK_HIP_VERSION_MINOR "${HIP_VERSION_MINOR}")
set(CK_HIP_VERSION_PATCH "${HIP_VERSION_PATCH}")
if( DEFINED CK_OVERRIDE_HIP_VERSION_MAJOR )
    set(CK_HIP_VERSION_MAJOR "${CK_OVERRIDE_HIP_VERSION_MAJOR}")
    message(STATUS "CK_HIP_VERSION_MAJOR overriden with ${CK_OVERRIDE_HIP_VERSION_MAJOR}")
endif()
if( DEFINED CK_OVERRIDE_HIP_VERSION_MINOR )
    set(CK_HIP_VERSION_MINOR "${CK_OVERRIDE_HIP_VERSION_MINOR}")
    message(STATUS "CK_HIP_VERSION_MINOR overriden with ${CK_OVERRIDE_HIP_VERSION_MINOR}")
endif()
if( DEFINED CK_OVERRIDE_HIP_VERSION_PATCH )
    set(CK_HIP_VERSION_PATCH "${CK_OVERRIDE_HIP_VERSION_PATCH}")
    message(STATUS "CK_HIP_VERSION_PATCH overriden with ${CK_OVERRIDE_HIP_VERSION_PATCH}")
endif()
message(STATUS "Build with HIP ${HIP_VERSION}")
361
link_libraries(hip::device)
362
363
364
365
366
if(CK_hip_VERSION VERSION_GREATER_EQUAL 6.0.23494)
    add_compile_definitions(__HIP_PLATFORM_AMD__=1)
else()
    add_compile_definitions(__HIP_PLATFORM_HCC__=1)
endif()
367

Chao Liu's avatar
Chao Liu committed
368
369
## tidy
include(EnableCompilerWarnings)
JD's avatar
JD committed
370
set(CK_TIDY_ERRORS ERRORS * -readability-inconsistent-declaration-parameter-name)
Chao Liu's avatar
Chao Liu committed
371
if(CMAKE_CXX_COMPILER MATCHES ".*hcc" OR CMAKE_CXX_COMPILER MATCHES ".*clang\\+\\+")
JD's avatar
JD committed
372
    set(CK_TIDY_CHECKS -modernize-use-override -readability-non-const-parameter)
Chao Liu's avatar
Chao Liu committed
373
# Enable tidy on hip
JD's avatar
JD committed
374
375
elseif(CK_BACKEND STREQUAL "HIP" OR CK_BACKEND STREQUAL "HIPNOGPU")
    set(CK_TIDY_ERRORS ALL)
Chao Liu's avatar
Chao Liu committed
376
377
endif()

JD's avatar
JD committed
378

Chao Liu's avatar
Chao Liu committed
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
include(ClangTidy)
enable_clang_tidy(
    CHECKS
        *
        -abseil-*
        -android-cloexec-fopen
        # Yea we shouldn't be using rand()
        -cert-msc30-c
        -bugprone-exception-escape
        -bugprone-macro-parentheses
        -cert-env33-c
        -cert-msc32-c
        -cert-msc50-cpp
        -cert-msc51-cpp
        -cert-dcl37-c
        -cert-dcl51-cpp
        -clang-analyzer-alpha.core.CastToStruct
        -clang-analyzer-optin.performance.Padding
        -clang-diagnostic-deprecated-declarations
        -clang-diagnostic-extern-c-compat
        -clang-diagnostic-unused-command-line-argument
        -cppcoreguidelines-avoid-c-arrays
        -cppcoreguidelines-avoid-magic-numbers
        -cppcoreguidelines-explicit-virtual-functions
        -cppcoreguidelines-init-variables
        -cppcoreguidelines-macro-usage
        -cppcoreguidelines-non-private-member-variables-in-classes
        -cppcoreguidelines-pro-bounds-array-to-pointer-decay
        -cppcoreguidelines-pro-bounds-constant-array-index
        -cppcoreguidelines-pro-bounds-pointer-arithmetic
        -cppcoreguidelines-pro-type-member-init
        -cppcoreguidelines-pro-type-reinterpret-cast
        -cppcoreguidelines-pro-type-union-access
        -cppcoreguidelines-pro-type-vararg
        -cppcoreguidelines-special-member-functions
        -fuchsia-*
        -google-explicit-constructor
        -google-readability-braces-around-statements
        -google-readability-todo
        -google-runtime-int
        -google-runtime-references
        -hicpp-vararg
        -hicpp-braces-around-statements
        -hicpp-explicit-conversions
        -hicpp-named-parameter
        -hicpp-no-array-decay
        # We really shouldn't use bitwise operators with signed integers, but
        # opencl leaves us no choice
        -hicpp-avoid-c-arrays
        -hicpp-signed-bitwise
        -hicpp-special-member-functions
        -hicpp-uppercase-literal-suffix
        -hicpp-use-auto
        -hicpp-use-equals-default
        -hicpp-use-override
        -llvm-header-guard
        -llvm-include-order
        #-llvmlibc-*
        -llvmlibc-restrict-system-libc-headers
        -llvmlibc-callee-namespace
        -llvmlibc-implementation-in-namespace
        -llvm-else-after-return
        -llvm-qualified-auto
        -misc-misplaced-const
        -misc-non-private-member-variables-in-classes
        -misc-no-recursion
        -modernize-avoid-bind
        -modernize-avoid-c-arrays
        -modernize-pass-by-value
        -modernize-use-auto
        -modernize-use-default-member-init
        -modernize-use-equals-default
        -modernize-use-trailing-return-type
        -modernize-use-transparent-functors
        -performance-unnecessary-value-param
        -readability-braces-around-statements
        -readability-else-after-return
        # we are not ready to use it, but very useful
        -readability-function-cognitive-complexity
        -readability-isolate-declaration
        -readability-magic-numbers
        -readability-named-parameter
        -readability-uppercase-literal-suffix
        -readability-convert-member-functions-to-static
        -readability-qualified-auto
        -readability-redundant-string-init
        # too many narrowing conversions in our code
        -bugprone-narrowing-conversions
        -cppcoreguidelines-narrowing-conversions
        -altera-struct-pack-align
        -cppcoreguidelines-prefer-member-initializer
JD's avatar
JD committed
470
471
        ${CK_TIDY_CHECKS}
        ${CK_TIDY_ERRORS}
Chao Liu's avatar
Chao Liu committed
472
473
474
    HEADER_FILTER
        "\.hpp$"
    EXTRA_ARGS
JD's avatar
JD committed
475
        -DCK_USE_CLANG_TIDY
Chao Liu's avatar
Chao Liu committed
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
)

include(CppCheck)
enable_cppcheck(
    CHECKS
        warning
        style
        performance
        portability
    SUPPRESS
        ConfigurationNotChecked
        constStatement
        duplicateCondition
        noExplicitConstructor
        passedByValue
Chao Liu's avatar
Chao Liu committed
491
        preprocessorErrorDirective
Chao Liu's avatar
Chao Liu committed
492
493
494
495
496
497
498
        shadowVariable
        unusedFunction
        unusedPrivateFunction
        unusedStructMember
        unmatchedSuppression
    FORCE
    SOURCES
Chao Liu's avatar
Chao Liu committed
499
        library/src
Chao Liu's avatar
Chao Liu committed
500
501
502
    INCLUDE
        ${CMAKE_CURRENT_SOURCE_DIR}/include
        ${CMAKE_CURRENT_BINARY_DIR}/include
Chao Liu's avatar
Chao Liu committed
503
        ${CMAKE_CURRENT_SOURCE_DIR}/library/include
Chao Liu's avatar
Chao Liu committed
504
505
506
507
    DEFINE
        CPPCHECK=1
        __linux__=1
)
Chao Liu's avatar
Chao Liu committed
508

JD's avatar
JD committed
509
510
511
512
set(CMAKE_LIBRARY_OUTPUT_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/lib)
set(CMAKE_ARCHIVE_OUTPUT_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/lib)
set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/bin)

513
# set CK project include directories
Chao Liu's avatar
Chao Liu committed
514
include_directories(BEFORE
515
    ${PROJECT_BINARY_DIR}/include
Chao Liu's avatar
Chao Liu committed
516
517
    ${PROJECT_SOURCE_DIR}/include
    ${PROJECT_SOURCE_DIR}/library/include
518
    ${HIP_INCLUDE_DIRS}
JD's avatar
JD committed
519
)
Chao Liu's avatar
Chao Liu committed
520

JD's avatar
JD committed
521
522
SET(BUILD_DEV ON CACHE BOOL "BUILD_DEV")
if(BUILD_DEV)
523
524
    add_compile_options(-Werror)
    add_compile_options(-Weverything)
JD's avatar
JD committed
525
526
527
endif()
message("CMAKE_CXX_FLAGS: ${CMAKE_CXX_FLAGS}")

528
529
530
531
532
533
534
if("${CMAKE_CXX_COMPILER_ID}" MATCHES "Clang")
    add_compile_options(-fcolor-diagnostics)
endif()
if("${CMAKE_CXX_COMPILER_ID}" STREQUAL "GNU" AND CMAKE_CXX_COMPILER_VERSION VERSION_GREATER 4.9)
    add_compile_options(-fdiagnostics-color=always)
endif()

Anthony Chang's avatar
Anthony Chang committed
535
536
add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C ${CMAKE_CFG_INTDIR})

537
538
539
540
file(GLOB_RECURSE INSTANCE_FILES "${PROJECT_SOURCE_DIR}/*/device_*_instance.cpp")
file(GLOB dir_list RELATIVE ${PROJECT_SOURCE_DIR}/library/src/tensor_operation_instance/gpu ${PROJECT_SOURCE_DIR}/library/src/tensor_operation_instance/gpu/*)
set(CK_DEVICE_INSTANCES)
FOREACH(subdir_path ${dir_list})
541
542
543
544
545
set(target_dir)
IF(IS_DIRECTORY "${PROJECT_SOURCE_DIR}/library/src/tensor_operation_instance/gpu/${subdir_path}")
    set(cmake_instance)
    file(READ "${PROJECT_SOURCE_DIR}/library/src/tensor_operation_instance/gpu/${subdir_path}/CMakeLists.txt" cmake_instance)
    set(add_inst 0)
546
    if(("${cmake_instance}" MATCHES "fp8" OR "${cmake_instance}" MATCHES "_f8") AND DTYPES MATCHES "fp8")
547
        set(add_inst 1)
548
    endif()
549
    if(("${cmake_instance}" MATCHES "bf8" OR "${cmake_instance}" MATCHES "_b8") AND DTYPES MATCHES "bf8")
550
551
        set(add_inst 1)
    endif()
552
    if(("${cmake_instance}" MATCHES "fp16" OR "${cmake_instance}" MATCHES "_f16") AND DTYPES MATCHES "fp16")
553
        set(add_inst 1)
554
    endif()
555
    if(("${cmake_instance}" MATCHES "fp32" OR "${cmake_instance}" MATCHES "_f32") AND DTYPES MATCHES "fp32")
556
        set(add_inst 1)
557
    endif()
558
    if(("${cmake_instance}" MATCHES "fp64" OR "${cmake_instance}" MATCHES "_f64") AND DTYPES MATCHES "fp64")
559
        set(add_inst 1)
560
    endif()
561
    if(("${cmake_instance}" MATCHES "bf16" OR "${cmake_instance}" MATCHES "_b16") AND DTYPES MATCHES "bf16")
562
        set(add_inst 1)
563
    endif()
564
    if(("${cmake_instance}" MATCHES "int8" OR "${cmake_instance}" MATCHES "_i8") AND DTYPES MATCHES "int8")
565
        set(add_inst 1)
566
567
    endif()
    if(NOT "${cmake_instance}" MATCHES "DTYPES")
568
        set(add_inst 1)
569
570
    endif()
    if(add_inst EQUAL 1 OR NOT DEFINED DTYPES)
571
        list(APPEND CK_DEVICE_INSTANCES device_${subdir_path}_instance)
572
573
    endif()
ENDIF()
574
ENDFOREACH()
575

576
add_custom_target(instances DEPENDS utility;${CK_DEVICE_INSTANCES}  SOURCES ${INSTANCE_FILES})
577
add_subdirectory(library)
578

579
if(NOT GPU_ARCHS AND USER_GPU_TARGETS)
580
   rocm_package_setup_component(tests
581
582
        LIBRARY_NAME composablekernel
        PACKAGE_NAME tests # Prevent -static suffix on package name
583
   )
584

585
   rocm_package_setup_component(examples
586
587
        LIBRARY_NAME composablekernel
        PACKAGE_NAME examples
588
   )
589
   add_subdirectory(example)
arai713's avatar
arai713 committed
590
591
592
   if(BUILD_TESTING)
	   add_subdirectory(test)
   endif()
593
endif()
JD's avatar
JD committed
594

595
596
597
598
599
600
rocm_package_setup_component(profiler
    LIBRARY_NAME composablekernel
    PACKAGE_NAME ckprofiler
)
add_subdirectory(profiler)

601
if(CK_USE_CODEGEN AND (SUPPORTED_GPU_TARGETS MATCHES "gfx9" OR GPU_ARCHS))
arai713's avatar
arai713 committed
602
603
604
  add_subdirectory(codegen)
endif()

JD's avatar
JD committed
605
606
607
608
609
610
611
612
613
#Create an interface target for the include only files and call it "composablekernels"
include(CMakePackageConfigHelpers)

write_basic_package_version_file(
    "${CMAKE_CURRENT_BINARY_DIR}/composable_kernelConfigVersion.cmake"
    VERSION "${version}"
    COMPATIBILITY AnyNewerVersion
)

Anthony Chang's avatar
Anthony Chang committed
614
configure_package_config_file(${CMAKE_CURRENT_SOURCE_DIR}/Config.cmake.in
615
616
617
    "${CMAKE_CURRENT_BINARY_DIR}/composable_kernelConfig.cmake"
    INSTALL_DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/composable_kernel
    NO_CHECK_REQUIRED_COMPONENTS_MACRO
JD's avatar
JD committed
618
619
)

620
rocm_install(FILES
JD's avatar
JD committed
621
622
    "${CMAKE_CURRENT_BINARY_DIR}/composable_kernelConfig.cmake"
    "${CMAKE_CURRENT_BINARY_DIR}/composable_kernelConfigVersion.cmake"
Anthony Chang's avatar
Anthony Chang committed
623
    DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/composable_kernel
JD's avatar
JD committed
624
)
625

626
# Install CK version and configuration files
627
rocm_install(FILES
628
629
630
631
632
    ${PROJECT_BINARY_DIR}/include/ck/version.h
    ${PROJECT_BINARY_DIR}/include/ck/config.h
    DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}/ck/
)

633
634
635
636
637
638
639
640
641
642
set(CPACK_RESOURCE_FILE_LICENSE "${CMAKE_CURRENT_SOURCE_DIR}/LICENSE")
set(CPACK_RPM_PACKAGE_LICENSE "MIT")

rocm_create_package(
    NAME composablekernel
    DESCRIPTION "High Performance Composable Kernel for AMD GPUs"
    MAINTAINER "MIOpen Kernels Dev Team <dl.MIOpen@amd.com>"
    LDCONFIG
    HEADER_ONLY
)