root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-25 02:44:36 +00:00
Code Issues Actions 9 Packages Projects Releases Wiki Activity

build: add armv{6,7,8} support to cmake (#1251)

Browse Source 
- flags copied from Makefile
- updated comments in both CMakeLists.txt and Makefile to match reality
This commit is contained in:
Pavol Rusnak 2023-04-30 20:48:38 +02:00 committed by GitHub
parent a5d30b1f53
commit 6f79699286
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 19 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
CMakeLists.txt
View File

@ -258,9 +258,22 @@ if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "arm" OR ${CMAKE_SYSTEM_PROCESSOR} MATCHES
# TODO: arm msvc? # TODO: arm msvc?
else() else()
if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "aarch64") if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "aarch64")
# Apple M1, M2, etc.
# Raspberry Pi 3, 4, Zero 2 (64-bit)
add_compile_options(-mcpu=native) add_compile_options(-mcpu=native)
endif() endif()
# TODO: armv6,7,8 version specific flags if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv6")
# Raspberry Pi 1, Zero
add_compile_options(-mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access)
endif()
if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv7")
# Raspberry Pi 2
add_compile_options(-mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access -funsafe-math-optimizations)
endif()
if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv8")
# Raspberry Pi 3, 4, Zero 2 (32-bit)
add_compile_options(-mfp16-format=ieee -mno-unaligned-access)
endif()
endif() endif()
elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "^(x86_64|i686|AMD64)$") elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "^(x86_64|i686|AMD64)$")
message(STATUS "x86 detected") message(STATUS "x86 detected")

8
Makefile
View File

@ -135,19 +135,21 @@ ifdef LLAMA_PERF
CXXFLAGS += -DGGML_PERF CXXFLAGS += -DGGML_PERF
endif endif
ifneq ($(filter aarch64%,$(UNAME_M)),) ifneq ($(filter aarch64%,$(UNAME_M)),)
# Apple M1, M2, etc.
# Raspberry Pi 3, 4, Zero 2 (64-bit)
CFLAGS += -mcpu=native CFLAGS += -mcpu=native
CXXFLAGS += -mcpu=native CXXFLAGS += -mcpu=native
endif endif
ifneq ($(filter armv6%,$(UNAME_M)),) ifneq ($(filter armv6%,$(UNAME_M)),)
# Raspberry Pi 1, 2, 3 # Raspberry Pi 1, Zero
CFLAGS += -mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access CFLAGS += -mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access
endif endif
ifneq ($(filter armv7%,$(UNAME_M)),) ifneq ($(filter armv7%,$(UNAME_M)),)
# Raspberry Pi 4 # Raspberry Pi 2
CFLAGS += -mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access -funsafe-math-optimizations CFLAGS += -mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access -funsafe-math-optimizations
endif endif
ifneq ($(filter armv8%,$(UNAME_M)),) ifneq ($(filter armv8%,$(UNAME_M)),)
# Raspberry Pi 4 # Raspberry Pi 3, 4, Zero 2 (32-bit)
CFLAGS += -mfp16-format=ieee -mno-unaligned-access CFLAGS += -mfp16-format=ieee -mno-unaligned-access
endif endif