CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-30 08:42:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml-cpu: add nnpa compile flag

Browse Source 
Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
(cherry picked from commit 4a9f60c201)
This commit is contained in:
Aaron Teo
2025-06-20 17:53:38 +08:00
parent bb16041cae
commit 5801806f70
5 changed files with 20 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
ggml/CMakeLists.txt
View File

@@ -131,6 +131,7 @@ option(GGML_RVV "ggml: enable rvv" ON)
option(GGML_RV_ZFH "ggml: enable riscv zfh" OFF) option(GGML_RV_ZFH "ggml: enable riscv zfh" OFF)
option(GGML_XTHEADVECTOR "ggml: enable xtheadvector" OFF) option(GGML_XTHEADVECTOR "ggml: enable xtheadvector" OFF)
option(GGML_VXE "ggml: enable vxe" ON) option(GGML_VXE "ggml: enable vxe" ON)
option(GGML_NNPA "ggml: enable nnpa" ON)
option(GGML_CPU_ALL_VARIANTS "ggml: build all variants of the CPU backend (requires GGML_BACKEND_DL)" OFF) option(GGML_CPU_ALL_VARIANTS "ggml: build all variants of the CPU backend (requires GGML_BACKEND_DL)" OFF)
set(GGML_CPU_ARM_ARCH "" CACHE STRING "ggml: CPU architecture for ARM") set(GGML_CPU_ARM_ARCH "" CACHE STRING "ggml: CPU architecture for ARM")

1
ggml/include/ggml-cpu.h
View File

@@ -101,6 +101,7 @@ extern "C" {
GGML_BACKEND_API int ggml_cpu_has_riscv_v (void); GGML_BACKEND_API int ggml_cpu_has_riscv_v (void);
GGML_BACKEND_API int ggml_cpu_has_vsx (void); GGML_BACKEND_API int ggml_cpu_has_vsx (void);
GGML_BACKEND_API int ggml_cpu_has_vxe (void); GGML_BACKEND_API int ggml_cpu_has_vxe (void);
GGML_BACKEND_API int ggml_cpu_has_nnpa (void);
GGML_BACKEND_API int ggml_cpu_has_wasm_simd (void); GGML_BACKEND_API int ggml_cpu_has_wasm_simd (void);
GGML_BACKEND_API int ggml_cpu_has_llamafile (void); GGML_BACKEND_API int ggml_cpu_has_llamafile (void);

7
ggml/src/ggml-cpu/CMakeLists.txt
View File

@@ -448,6 +448,7 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
# TODO: Separation to determine activation of VX/VXE/VXE2 # TODO: Separation to determine activation of VX/VXE/VXE2
if (${S390X_M} MATCHES "8561|8562") if (${S390X_M} MATCHES "8561|8562")
set(GGML_NNPA OFF)
message(STATUS "z15 target") message(STATUS "z15 target")
list(APPEND ARCH_FLAGS -march=z15) list(APPEND ARCH_FLAGS -march=z15)
elseif (${S390X_M} MATCHES "3931") elseif (${S390X_M} MATCHES "3931")
@@ -464,8 +465,14 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
endif() endif()
if (GGML_VXE) if (GGML_VXE)
message(STATUS "VX/VXE/VXE2 enabled")
list(APPEND ARCH_FLAGS -mvx -mzvector) list(APPEND ARCH_FLAGS -mvx -mzvector)
endif() endif()
if (GGML_NNPA)
target_compile_definitions(${GGML_CPU_NAME} PRIVATE GGML_NNPA)
message(STATUS "NNPA enabled")
endif()
elseif (CMAKE_SYSTEM_PROCESSOR MATCHES "wasm") elseif (CMAKE_SYSTEM_PROCESSOR MATCHES "wasm")
message(STATUS "Wasm detected") message(STATUS "Wasm detected")
list (APPEND GGML_CPU_SOURCES ggml-cpu/arch/wasm/quants.c) list (APPEND GGML_CPU_SOURCES ggml-cpu/arch/wasm/quants.c)

8
ggml/src/ggml-cpu/ggml-cpu.c
View File

@@ -3369,6 +3369,14 @@ int ggml_cpu_has_vxe(void) {
#endif #endif
} }
int ggml_cpu_has_nnpa(void) {
#if defined(GGML_NNPA)
return 1;
#else
return 0;
#endif
}
int ggml_cpu_has_neon(void) { int ggml_cpu_has_neon(void) {
#if defined(__ARM_ARCH) && defined(__ARM_NEON) #if defined(__ARM_ARCH) && defined(__ARM_NEON)
return 1; return 1;

3
ggml/src/ggml-cpu/ggml-cpu.cpp
View File

@@ -578,6 +578,9 @@ static ggml_backend_feature * ggml_backend_cpu_get_features(ggml_backend_reg_t r
if (ggml_cpu_has_vxe()) { if (ggml_cpu_has_vxe()) {
features.push_back({ "VXE", "1" }); features.push_back({ "VXE", "1" });
} }
if (ggml_cpu_has_nnpa()) {
features.push_back({ "NNPA", "1" });
}
if (ggml_cpu_has_wasm_simd()) { if (ggml_cpu_has_wasm_simd()) {
features.push_back({ "WASM_SIMD", "1" }); features.push_back({ "WASM_SIMD", "1" });
} }