ggml-cpu : split arch-specific implementations (#13892)
* move ggml-cpu-aarch64 to repack * split quantize_row_q8_0/1 * split helper functions * split ggml_vec_dot_q4_0_q8_0 * split ggml_vec_dot_q4_1_q8_1 * split ggml_vec_dot_q5_0_q8_0 * split ggml_vec_dot_q5_1_q8_1 * split ggml_vec_dot_q8_0_q8_0 * split ggml_vec_dot_tq1_0_q8_K * split ggml_vec_dot_tq2_0_q8_K * split ggml_vec_dot_q2_K_q8_K * split ggml_vec_dot_q3_K_q8_K * split ggml_vec_dot_q4_K_q8_K * split ggml_vec_dot_q5_K_q8_K * split ggml_vec_dot_q6_K_q8_K * split ggml_vec_dot_iq2_xxs_q8_K * split ggml_vec_dot_iq2_xs_q8_K * split ggml_vec_dot_iq2_s_q8_K * split ggml_vec_dot_iq3_xxs_q8_K * split ggml_vec_dot_iq3_s_q8_K * split ggml_vec_dot_iq1_s_q8_K * split ggml_vec_dot_iq1_m_q8_K * split ggml_vec_dot_iq4_nl_q8_0 * split ggml_vec_dot_iq4_xs_q8_K * fix typos * fix missing prototypes * rename ggml-cpu-quants.c * rename ggml-cpu-traits * rename arm folder * move cpu-feats-x86.cpp * rename ggml-cpu-hbm * update arm detection macro in quants.c * move iq quant tables * split ggml_quantize_mat_q8_0/K * split ggml_gemv_* * split ggml_gemm_* * rename namespace aarch64 to repack * use weak aliases to replace test macros * rename GGML_CPU_AARCH64 to GGML_CPU_REPACK * rename more aarch64 to repack * clean up rebase leftover * fix compilation errors * remove trailing spaces * try to fix clang compilation errors * try to fix clang compilation errors again * try to fix clang compilation errors, 3rd attempt * try to fix clang compilation errors, 4th attempt * try to fix clang compilation errors, 5th attempt * try to fix clang compilation errors, 6th attempt * try to fix clang compilation errors, 7th attempt * try to fix clang compilation errors, 8th attempt * try to fix clang compilation errors, 9th attempt * more cleanup * fix compilation errors * fix apple targets * fix a typo in arm version of ggml_vec_dot_q4_K_q8_K Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
xctan
GitHub
parent
8f47e25f56
commit
f470bc36be
36 changed files with 24288 additions and 17167 deletions
|
|
@ -10,14 +10,14 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
|||
list (APPEND GGML_CPU_SOURCES
|
||||
ggml-cpu/ggml-cpu.c
|
||||
ggml-cpu/ggml-cpu.cpp
|
||||
ggml-cpu/ggml-cpu-aarch64.cpp
|
||||
ggml-cpu/ggml-cpu-aarch64.h
|
||||
ggml-cpu/ggml-cpu-hbm.cpp
|
||||
ggml-cpu/ggml-cpu-hbm.h
|
||||
ggml-cpu/ggml-cpu-quants.c
|
||||
ggml-cpu/ggml-cpu-quants.h
|
||||
ggml-cpu/ggml-cpu-traits.cpp
|
||||
ggml-cpu/ggml-cpu-traits.h
|
||||
ggml-cpu/repack.cpp
|
||||
ggml-cpu/repack.h
|
||||
ggml-cpu/hbm.cpp
|
||||
ggml-cpu/hbm.h
|
||||
ggml-cpu/quants.c
|
||||
ggml-cpu/quants.h
|
||||
ggml-cpu/traits.cpp
|
||||
ggml-cpu/traits.h
|
||||
ggml-cpu/amx/amx.cpp
|
||||
ggml-cpu/amx/amx.h
|
||||
ggml-cpu/amx/mmq.cpp
|
||||
|
|
@ -84,6 +84,11 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
|||
|
||||
if (GGML_SYSTEM_ARCH STREQUAL "ARM")
|
||||
message(STATUS "ARM detected")
|
||||
list(APPEND GGML_CPU_SOURCES
|
||||
ggml-cpu/arch/arm/quants.c
|
||||
ggml-cpu/arch/arm/repack.cpp
|
||||
)
|
||||
|
||||
if (MSVC AND NOT CMAKE_C_COMPILER_ID STREQUAL "Clang")
|
||||
message(FATAL_ERROR "MSVC is not supported for ARM, use clang")
|
||||
else()
|
||||
|
|
@ -167,6 +172,11 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
|||
endif()
|
||||
elseif (GGML_SYSTEM_ARCH STREQUAL "x86")
|
||||
message(STATUS "x86 detected")
|
||||
list(APPEND GGML_CPU_SOURCES
|
||||
ggml-cpu/arch/x86/quants.c
|
||||
ggml-cpu/arch/x86/repack.cpp
|
||||
)
|
||||
|
||||
if (MSVC)
|
||||
# instruction set detection for MSVC only
|
||||
if (GGML_NATIVE)
|
||||
|
|
@ -302,7 +312,7 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
|||
# Since multiple variants of the CPU backend may be included in the same
|
||||
# build, using set_source_files_properties() to set the arch flags is not possible
|
||||
set(GGML_CPU_FEATS_NAME ${GGML_CPU_NAME}-feats)
|
||||
add_library(${GGML_CPU_FEATS_NAME} OBJECT ggml-cpu/cpu-feats-x86.cpp)
|
||||
add_library(${GGML_CPU_FEATS_NAME} OBJECT ggml-cpu/arch/x86/cpu-feats.cpp)
|
||||
target_include_directories(${GGML_CPU_FEATS_NAME} PRIVATE . .. ../include)
|
||||
target_compile_definitions(${GGML_CPU_FEATS_NAME} PRIVATE ${ARCH_DEFINITIONS})
|
||||
target_compile_definitions(${GGML_CPU_FEATS_NAME} PRIVATE GGML_BACKEND_DL GGML_BACKEND_BUILD GGML_BACKEND_SHARED)
|
||||
|
|
@ -311,6 +321,7 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
|||
endif()
|
||||
elseif (GGML_SYSTEM_ARCH STREQUAL "PowerPC")
|
||||
message(STATUS "PowerPC detected")
|
||||
list(APPEND GGML_CPU_SOURCES ggml-cpu/arch/powerpc/quants.c)
|
||||
if (GGML_NATIVE)
|
||||
if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "ppc64")
|
||||
file(READ "/proc/cpuinfo" POWER10_M)
|
||||
|
|
@ -338,6 +349,8 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
|||
endif()
|
||||
elseif (GGML_SYSTEM_ARCH STREQUAL "loongarch64")
|
||||
message(STATUS "loongarch64 detected")
|
||||
list(APPEND GGML_CPU_SOURCES ggml-cpu/arch/loongarch/quants.c)
|
||||
|
||||
list(APPEND ARCH_FLAGS -march=loongarch64)
|
||||
if (GGML_LASX)
|
||||
list(APPEND ARCH_FLAGS -mlasx)
|
||||
|
|
@ -347,6 +360,10 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
|||
endif()
|
||||
elseif (GGML_SYSTEM_ARCH STREQUAL "riscv64")
|
||||
message(STATUS "riscv64 detected")
|
||||
list(APPEND GGML_CPU_SOURCES
|
||||
ggml-cpu/arch/riscv/quants.c
|
||||
ggml-cpu/arch/riscv/repack.cpp
|
||||
)
|
||||
if (GGML_RVV)
|
||||
if (GGML_XTHEADVECTOR)
|
||||
list(APPEND ARCH_FLAGS -march=rv64gc_xtheadvector -mabi=lp64d)
|
||||
|
|
@ -358,6 +375,7 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
|||
endif()
|
||||
elseif (GGML_SYSTEM_ARCH STREQUAL "s390x")
|
||||
message(STATUS "s390x detected")
|
||||
list(APPEND GGML_CPU_SOURCES ggml-cpu/arch/s390/quants.c)
|
||||
file(READ "/proc/cpuinfo" CPUINFO_CONTENTS)
|
||||
string(REGEX REPLACE "machine[ \t\r\n]*=[ \t\r\n]*([0-9]+)" "\\1" S390X_M ${CPUINFO_CONTENTS})
|
||||
|
||||
|
|
@ -381,12 +399,16 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
|||
if (GGML_VXE)
|
||||
list(APPEND ARCH_FLAGS -mvx -mzvector)
|
||||
endif()
|
||||
elseif (CMAKE_SYSTEM_PROCESSOR MATCHES "wasm")
|
||||
message(STATUS "Wasm detected")
|
||||
list (APPEND GGML_CPU_SOURCES ggml-cpu/arch/wasm/quants.c)
|
||||
else()
|
||||
message(STATUS "Unknown architecture")
|
||||
message(WARNING "Unknown CPU architecture. Falling back to generic implementations.")
|
||||
list(APPEND ARCH_FLAGS -DGGML_CPU_GENERIC)
|
||||
endif()
|
||||
|
||||
if (GGML_CPU_AARCH64)
|
||||
target_compile_definitions(${GGML_CPU_NAME} PRIVATE GGML_USE_CPU_AARCH64)
|
||||
if (GGML_CPU_REPACK)
|
||||
target_compile_definitions(${GGML_CPU_NAME} PRIVATE GGML_USE_CPU_REPACK)
|
||||
endif()
|
||||
|
||||
if (GGML_CPU_KLEIDIAI)
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue