CMakeLists.txt 10.1 KB
Newer Older
xuxzh1's avatar
init  
xuxzh1 committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
cmake_minimum_required(VERSION 3.14) # for add_link_options and implicit target directories.
project("ggml" C CXX)
include(CheckIncludeFileCXX)

set(CMAKE_EXPORT_COMPILE_COMMANDS ON)

if (NOT XCODE AND NOT MSVC AND NOT CMAKE_BUILD_TYPE)
    set(CMAKE_BUILD_TYPE Release CACHE STRING "Build type" FORCE)
    set_property(CACHE CMAKE_BUILD_TYPE PROPERTY STRINGS "Debug" "Release" "MinSizeRel" "RelWithDebInfo")
endif()

if (CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR)
    set(GGML_STANDALONE ON)

    set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)

    # configure project version
    # TODO
else()
    set(GGML_STANDALONE OFF)
endif()

if (EMSCRIPTEN)
    set(BUILD_SHARED_LIBS_DEFAULT OFF)

    option(GGML_WASM_SINGLE_FILE "ggml: embed WASM inside the generated ggml.js" ON)
else()
    if (MINGW)
        set(BUILD_SHARED_LIBS_DEFAULT OFF)
    else()
        set(BUILD_SHARED_LIBS_DEFAULT ON)
    endif()
endif()

option(BUILD_SHARED_LIBS "ggml: build shared libraries" ${BUILD_SHARED_LIBS_DEFAULT})

#
# option list
#

# TODO: mark all options as advanced when not GGML_STANDALONE

if (APPLE)
    set(GGML_METAL_DEFAULT ON)
    set(GGML_BLAS_DEFAULT ON)
    set(GGML_BLAS_VENDOR_DEFAULT "Apple")
else()
    set(GGML_METAL_DEFAULT OFF)
    set(GGML_BLAS_DEFAULT OFF)
    set(GGML_BLAS_VENDOR_DEFAULT "Generic")
endif()

if (CMAKE_CROSSCOMPILING)
    set(GGML_NATIVE_DEFAULT OFF)
else()
    set(GGML_NATIVE_DEFAULT ON)
endif()

xuxzh1's avatar
update  
xuxzh1 committed
59
60
61
62
63
64
65
66
67
# defaults
if (NOT GGML_LLAMAFILE_DEFAULT)
    set(GGML_LLAMAFILE_DEFAULT OFF)
endif()

if (NOT GGML_CUDA_GRAPHS_DEFAULT)
    set(GGML_CUDA_GRAPHS_DEFAULT OFF)
endif()

xuxzh1's avatar
init  
xuxzh1 committed
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
# general
option(GGML_STATIC "ggml: static link libraries"         OFF)
option(GGML_NATIVE "ggml: enable -march=native flag"     ${GGML_NATIVE_DEFAULT})
option(GGML_LTO    "ggml: enable link time optimization" OFF)
option(GGML_CCACHE "ggml: use ccache if available"       ON)

# debug
option(GGML_ALL_WARNINGS           "ggml: enable all compiler warnings"                   ON)
option(GGML_ALL_WARNINGS_3RD_PARTY "ggml: enable all compiler warnings in 3rd party libs" OFF)
option(GGML_GPROF                  "ggml: enable gprof"                                   OFF)

# build
option(GGML_FATAL_WARNINGS    "ggml: enable -Werror flag"    OFF)

# sanitizers
option(GGML_SANITIZE_THREAD    "ggml: enable thread sanitizer"    OFF)
option(GGML_SANITIZE_ADDRESS   "ggml: enable address sanitizer"   OFF)
option(GGML_SANITIZE_UNDEFINED "ggml: enable undefined sanitizer" OFF)

# instruction set specific
if (GGML_NATIVE OR NOT GGML_NATIVE_DEFAULT)
    set(INS_ENB OFF)
else()
    set(INS_ENB ON)
endif()

option(GGML_CPU_HBM     "ggml: use memkind for CPU HBM" OFF)
xuxzh1's avatar
update  
xuxzh1 committed
95
option(GGML_CPU_AARCH64 "ggml: use runtime weight conversion of Q4_0 to Q4_X_X" ON)
xuxzh1's avatar
init  
xuxzh1 committed
96
97
98
99
100
101
102

option(GGML_AVX         "ggml: enable AVX"              ${INS_ENB})
option(GGML_AVX2        "ggml: enable AVX2"             ${INS_ENB})
option(GGML_AVX512      "ggml: enable AVX512"           OFF)
option(GGML_AVX512_VBMI "ggml: enable AVX512-VBMI"      OFF)
option(GGML_AVX512_VNNI "ggml: enable AVX512-VNNI"      OFF)
option(GGML_AVX512_BF16 "ggml: enable AVX512-BF16"      OFF)
xuxzh1's avatar
update  
xuxzh1 committed
103
104
105
option(GGML_AMX_TILE    "ggml: enable AMX-TILE"         OFF)
option(GGML_AMX_INT8    "ggml: enable AMX-INT8"         OFF)
option(GGML_AMX_BF16    "ggml: enable AMX-BF16"         OFF)
xuxzh1's avatar
init  
xuxzh1 committed
106
107
108
109
110
111
option(GGML_FMA         "ggml: enable FMA"              ${INS_ENB})
if (NOT MSVC)
    option(GGML_F16C    "ggml: enable F16C"             ${INS_ENB}) # in MSVC F16C is implied with AVX2/AVX512
endif()
option(GGML_LASX        "ggml: enable lasx"             ON)
option(GGML_LSX         "ggml: enable lsx"              ON)
xuxzh1's avatar
update  
xuxzh1 committed
112
option(GGML_RVV         "ggml: enable rvv"              ON)
xuxzh1's avatar
init  
xuxzh1 committed
113
114
115
116
117
118
119
120
option(GGML_SVE         "ggml: enable SVE"              OFF)

if (WIN32)
    set(GGML_WIN_VER "0x602" CACHE STRING "ggml: Windows Version")
endif()

# ggml core
set(GGML_SCHED_MAX_COPIES  "4" CACHE STRING "ggml: max input copies for pipeline parallelism")
xuxzh1's avatar
update  
xuxzh1 committed
121
option(GGML_CPU                             "ggml: enable CPU backend"                        ON)
xuxzh1's avatar
init  
xuxzh1 committed
122
123
124
125
126
127

# 3rd party libs / backends
option(GGML_ACCELERATE                      "ggml: enable Accelerate framework"               ON)
option(GGML_BLAS                            "ggml: use BLAS"                                  ${GGML_BLAS_DEFAULT})
set(GGML_BLAS_VENDOR ${GGML_BLAS_VENDOR_DEFAULT} CACHE STRING
                                            "ggml: BLAS library vendor")
xuxzh1's avatar
update  
xuxzh1 committed
128
option(GGML_LLAMAFILE                       "ggml: use LLAMAFILE"                             ${GGML_LLAMAFILE_DEFAULT})
xuxzh1's avatar
init  
xuxzh1 committed
129
130
131
132
133
134
135
136
137
138
139

option(GGML_CUDA                            "ggml: use CUDA"                                  OFF)
option(GGML_MUSA                            "ggml: use MUSA"                                  OFF)
option(GGML_CUDA_FORCE_MMQ                  "ggml: use mmq kernels instead of cuBLAS"         OFF)
option(GGML_CUDA_FORCE_CUBLAS               "ggml: always use cuBLAS instead of mmq kernels"  OFF)
option(GGML_CUDA_F16                        "ggml: use 16 bit floats for some calculations"   OFF)
set   (GGML_CUDA_PEER_MAX_BATCH_SIZE "128" CACHE STRING
                                            "ggml: max. batch size for using peer access")
option(GGML_CUDA_NO_PEER_COPY               "ggml: do not use peer to peer copies"            OFF)
option(GGML_CUDA_NO_VMM                     "ggml: do not try to use CUDA VMM"                OFF)
option(GGML_CUDA_FA_ALL_QUANTS              "ggml: compile all quants for FlashAttention"     OFF)
xuxzh1's avatar
update  
xuxzh1 committed
140
option(GGML_CUDA_GRAPHS                     "ggml: use CUDA graphs (llama.cpp only)"          ${GGML_CUDA_GRAPHS_DEFAULT})
xuxzh1's avatar
init  
xuxzh1 committed
141

xuxzh1's avatar
update  
xuxzh1 committed
142
option(GGML_HIP                             "ggml: use HIP"                                   OFF)
xuxzh1's avatar
init  
xuxzh1 committed
143
144
145
146
147
option(GGML_HIP_UMA                         "ggml: use HIP unified memory architecture"       OFF)
option(GGML_VULKAN                          "ggml: use Vulkan"                                OFF)
option(GGML_VULKAN_CHECK_RESULTS            "ggml: run Vulkan op checks"                      OFF)
option(GGML_VULKAN_DEBUG                    "ggml: enable Vulkan debug output"                OFF)
option(GGML_VULKAN_MEMORY_DEBUG             "ggml: enable Vulkan memory debug output"         OFF)
xuxzh1's avatar
update  
xuxzh1 committed
148
149
option(GGML_VULKAN_SHADER_DEBUG_INFO        "ggml: enable Vulkan shader debug info"           OFF)
option(GGML_VULKAN_PERF                     "ggml: enable Vulkan perf output"                 OFF)
xuxzh1's avatar
init  
xuxzh1 committed
150
151
152
153
option(GGML_VULKAN_VALIDATE                 "ggml: enable Vulkan validation"                  OFF)
option(GGML_VULKAN_RUN_TESTS                "ggml: run Vulkan tests"                          OFF)
option(GGML_KOMPUTE                         "ggml: use Kompute"                               OFF)
option(GGML_METAL                           "ggml: use Metal"                                 ${GGML_METAL_DEFAULT})
xuxzh1's avatar
update  
xuxzh1 committed
154
option(GGML_METAL_USE_BF16                  "ggml: use bfloat if available"                   OFF)
xuxzh1's avatar
init  
xuxzh1 committed
155
156
157
158
159
160
161
162
option(GGML_METAL_NDEBUG                    "ggml: disable Metal debugging"                   OFF)
option(GGML_METAL_SHADER_DEBUG              "ggml: compile Metal with -fno-fast-math"         OFF)
option(GGML_METAL_EMBED_LIBRARY             "ggml: embed Metal library"                       ${GGML_METAL})
set   (GGML_METAL_MACOSX_VERSION_MIN "" CACHE STRING
                                            "ggml: metal minimum macOS version")
set   (GGML_METAL_STD "" CACHE STRING       "ggml: metal standard version (-std flag)")
option(GGML_OPENMP                          "ggml: use OpenMP"                                ON)
option(GGML_RPC                             "ggml: use RPC"                                   OFF)
xuxzh1's avatar
update  
xuxzh1 committed
163
option(GGML_AMX                             "ggml: use AMX"                                   OFF)
xuxzh1's avatar
init  
xuxzh1 committed
164
165
166
167
option(GGML_SYCL                            "ggml: use SYCL"                                  OFF)
option(GGML_SYCL_F16                        "ggml: use 16 bit floats for sycl calculations"   OFF)
set   (GGML_SYCL_TARGET "INTEL" CACHE STRING
                                            "ggml: sycl target device")
xuxzh1's avatar
update  
xuxzh1 committed
168
169
set   (GGML_SYCL_DEVICE_ARCH "" CACHE STRING
                                            "ggml: sycl device architecture")
xuxzh1's avatar
init  
xuxzh1 committed
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221

# extra artifacts
option(GGML_BUILD_TESTS    "ggml: build tests"    ${GGML_STANDALONE})
option(GGML_BUILD_EXAMPLES "ggml: build examples" ${GGML_STANDALONE})

#
# dependencies
#

set(CMAKE_C_STANDARD 11)
set(CMAKE_C_STANDARD_REQUIRED true)

if (GGML_SYCL)
    set(CMAKE_CXX_STANDARD 17)
else()
    set(CMAKE_CXX_STANDARD 11)
endif()
set(CMAKE_CXX_STANDARD_REQUIRED true)

set(THREADS_PREFER_PTHREAD_FLAG ON)

find_package(Threads REQUIRED)

#
# build the library
#

add_subdirectory(src)

#
# tests and examples
#

if (GGML_BUILD_TESTS)
    enable_testing()
    add_subdirectory(tests)
endif ()

if (GGML_BUILD_EXAMPLES)
    add_subdirectory(examples)
endif ()

#
# install
#

include(GNUInstallDirs)
include(CMakePackageConfigHelpers)

# all public headers
set(GGML_PUBLIC_HEADERS
    include/ggml.h
xuxzh1's avatar
update  
xuxzh1 committed
222
    include/ggml-cpu.h
xuxzh1's avatar
init  
xuxzh1 committed
223
224
225
226
227
228
    include/ggml-alloc.h
    include/ggml-backend.h
    include/ggml-blas.h
    include/ggml-cann.h
    include/ggml-cuda.h
    include/ggml-kompute.h
xuxzh1's avatar
update  
xuxzh1 committed
229
    include/ggml-opt.h
xuxzh1's avatar
init  
xuxzh1 committed
230
231
232
233
234
235
236
237
238
    include/ggml-metal.h
    include/ggml-rpc.h
    include/ggml-sycl.h
    include/ggml-vulkan.h)

set_target_properties(ggml PROPERTIES PUBLIC_HEADER "${GGML_PUBLIC_HEADERS}")
#if (GGML_METAL)
#    set_target_properties(ggml PROPERTIES RESOURCE "${CMAKE_CURRENT_SOURCE_DIR}/src/ggml-metal.metal")
#endif()
xuxzh1's avatar
update  
xuxzh1 committed
239
240
install(TARGETS ggml LIBRARY PUBLIC_HEADER)
install(TARGETS ggml-base LIBRARY)
xuxzh1's avatar
init  
xuxzh1 committed
241

xuxzh1's avatar
update  
xuxzh1 committed
242
# FIXME: this should be done in the backend cmake files
xuxzh1's avatar
init  
xuxzh1 committed
243
if (GGML_METAL)
xuxzh1's avatar
update  
xuxzh1 committed
244
    # FIXME: does this need to be installed with GGML_METAL_EMBED_LIBRARY?
xuxzh1's avatar
init  
xuxzh1 committed
245
    install(
xuxzh1's avatar
update  
xuxzh1 committed
246
        FILES src/ggml-metal/ggml-metal.metal
xuxzh1's avatar
init  
xuxzh1 committed
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
        PERMISSIONS
            OWNER_READ
            OWNER_WRITE
            GROUP_READ
            WORLD_READ
        DESTINATION ${CMAKE_INSTALL_BINDIR})

    if (NOT GGML_METAL_EMBED_LIBRARY)
        install(
            FILES ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib
            DESTINATION ${CMAKE_INSTALL_BINDIR}
        )
    endif()
endif()

if (GGML_STANDALONE)
    configure_file(${CMAKE_CURRENT_SOURCE_DIR}/ggml.pc.in
        ${CMAKE_CURRENT_BINARY_DIR}/ggml.pc
        @ONLY)

    install(FILES ${CMAKE_CURRENT_BINARY_DIR}/ggml.pc
        DESTINATION share/pkgconfig)
endif()