kleidiai: add CPU feature detection to CI run script

[martin-klacer-arm]

This patch adds CPU feature detection for KleidiAI build. Previously, the -march flags for KleidiAI build were decided only by the present compiler capabilities without considering the current CPU, which is addressed in this patch.

[martin-klacer-arm]

Hi, just wondering if there's any updates on this? I'd appreciate another look if you have the time, thank you

[ggerganov]

Could you try to utilize and extend (if needed) the existing GGML_CPU_ALL_VARIANTS functionality to support this?

llama.cpp/ggml/src/CMakeLists.txt

Lines 351 to 447 in 29b28a9

	ggml_add_backend(CPU)
	if (GGML_CPU_ALL_VARIANTS)
	if (NOT GGML_BACKEND_DL)
	message(FATAL_ERROR "GGML_CPU_ALL_VARIANTS requires GGML_BACKEND_DL")
	elseif (GGML_CPU_ARM_ARCH)
	message(FATAL_ERROR "Cannot use both GGML_CPU_ARM_ARCH and GGML_CPU_ALL_VARIANTS")
	endif()
	if (GGML_SYSTEM_ARCH STREQUAL "x86")
	ggml_add_cpu_backend_variant(x64)
	ggml_add_cpu_backend_variant(sse42 SSE42)
	ggml_add_cpu_backend_variant(sandybridge SSE42 AVX)
	if (NOT MSVC)
	# __FMA__ and __F16C__ are not defined in MSVC, however they are implied with AVX2/AVX512
	ggml_add_cpu_backend_variant(ivybridge SSE42 AVX F16C)
	ggml_add_cpu_backend_variant(piledriver SSE42 AVX F16C FMA)
	endif()
	ggml_add_cpu_backend_variant(haswell SSE42 AVX F16C FMA AVX2 BMI2)
	ggml_add_cpu_backend_variant(skylakex SSE42 AVX F16C FMA AVX2 BMI2 AVX512)
	ggml_add_cpu_backend_variant(cannonlake SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VBMI)
	ggml_add_cpu_backend_variant(cascadelake SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VNNI)
	ggml_add_cpu_backend_variant(icelake SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VBMI AVX512_VNNI)
	if (NOT MSVC)
	# MSVC 2022 doesn't support BF16 intrinsics without `/arch:AVX10.1` ?!
	# https://learn.microsoft.com/en-us/cpp/intrinsics/x64-amd64-intrinsics-list?view=msvc-170
	# https://learn.microsoft.com/en-us/cpp/build/reference/arch-x64?view=msvc-170
	ggml_add_cpu_backend_variant(cooperlake SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VNNI AVX512_BF16)
	ggml_add_cpu_backend_variant(zen4 SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VBMI AVX512_VNNI AVX512_BF16)
	endif()
	ggml_add_cpu_backend_variant(alderlake SSE42 AVX F16C FMA AVX2 BMI2 AVX_VNNI)
	if (NOT MSVC)
	# MSVC doesn't support AMX
	ggml_add_cpu_backend_variant(sapphirerapids SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VBMI AVX512_VNNI AVX512_BF16 AMX_TILE AMX_INT8)
	endif()
	elseif(GGML_SYSTEM_ARCH STREQUAL "ARM")
	if (CMAKE_SYSTEM_NAME MATCHES "Linux")
	# Many of these features are optional so we build versions with popular
	# combinations and name the backends based on the version they were
	# first released with
	ggml_add_cpu_backend_variant(armv8.0_1)
	ggml_add_cpu_backend_variant(armv8.2_1 DOTPROD)
	ggml_add_cpu_backend_variant(armv8.2_2 DOTPROD FP16_VECTOR_ARITHMETIC)
	ggml_add_cpu_backend_variant(armv8.2_3 DOTPROD FP16_VECTOR_ARITHMETIC SVE)
	ggml_add_cpu_backend_variant(armv8.6_1 DOTPROD FP16_VECTOR_ARITHMETIC SVE MATMUL_INT8)
	ggml_add_cpu_backend_variant(armv8.6_2 DOTPROD FP16_VECTOR_ARITHMETIC SVE MATMUL_INT8 SVE2)
	ggml_add_cpu_backend_variant(armv9.2_1 DOTPROD FP16_VECTOR_ARITHMETIC SVE MATMUL_INT8 SME)
	ggml_add_cpu_backend_variant(armv9.2_2 DOTPROD FP16_VECTOR_ARITHMETIC SVE MATMUL_INT8 SVE2 SME)
	elseif (CMAKE_SYSTEM_NAME MATCHES "Android")
	# Android-specific backends with SoC-compatible feature sets
	ggml_add_cpu_backend_variant(android_armv8.0_1)
	ggml_add_cpu_backend_variant(android_armv8.2_1 DOTPROD)
	ggml_add_cpu_backend_variant(android_armv8.2_2 DOTPROD FP16_VECTOR_ARITHMETIC)
	ggml_add_cpu_backend_variant(android_armv8.6_1 DOTPROD FP16_VECTOR_ARITHMETIC MATMUL_INT8)
	ggml_add_cpu_backend_variant(android_armv9.0_1 DOTPROD MATMUL_INT8 FP16_VECTOR_ARITHMETIC SVE2)
	ggml_add_cpu_backend_variant(android_armv9.2_1 DOTPROD MATMUL_INT8 FP16_VECTOR_ARITHMETIC SVE SME)
	ggml_add_cpu_backend_variant(android_armv9.2_2 DOTPROD MATMUL_INT8 FP16_VECTOR_ARITHMETIC SVE SVE2 SME)
	elseif (APPLE)
	ggml_add_cpu_backend_variant(apple_m1 DOTPROD)
	ggml_add_cpu_backend_variant(apple_m2_m3 DOTPROD MATMUL_INT8)
	ggml_add_cpu_backend_variant(apple_m4 DOTPROD MATMUL_INT8 NOSVE SME)
	else()
	message(FATAL_ERROR "Unsupported ARM target OS: ${CMAKE_SYSTEM_NAME}")
	endif()
	elseif (GGML_SYSTEM_ARCH STREQUAL "PowerPC")
	if (CMAKE_SYSTEM_NAME MATCHES "Linux")
	ggml_add_cpu_backend_variant(power0)
	ggml_add_cpu_backend_variant(power7_1 POWER7)
	ggml_add_cpu_backend_variant(power7_2 POWER7 VSX)
	ggml_add_cpu_backend_variant(power8_1 POWER8)
	ggml_add_cpu_backend_variant(power8_2 POWER8 VSX)
	ggml_add_cpu_backend_variant(power9 POWER9 VSX)
	ggml_add_cpu_backend_variant(power10 POWER10 VSX)
	ggml_add_cpu_backend_variant(power11 POWER11 VSX)
	else()
	message(FATAL_ERROR "Unsupported PowerPC target OS: ${CMAKE_SYSTEM_NAME}")
	endif()
	elseif (GGML_SYSTEM_ARCH STREQUAL "s390x")
	if (CMAKE_SYSTEM_NAME MATCHES "Linux")
	ggml_add_cpu_backend_variant(z15 Z15 VXE2)
	ggml_add_cpu_backend_variant(z16 Z16 VXE2 NNPA)
	else()
	message(FATAL_ERROR "Unsupported s390x target OS: ${CMAKE_SYSTEM_NAME}")
	endif()
	elseif (GGML_SYSTEM_ARCH STREQUAL "riscv64")
	if (CMAKE_SYSTEM_NAME MATCHES "Linux")
	ggml_add_cpu_backend_variant(riscv64_0)
	ggml_add_cpu_backend_variant(riscv64_v RVV)
	else()
	message(FATAL_ERROR "Unsupported RISC-V target OS: ${CMAKE_SYSTEM_NAME}")
	endif()
	else()
	message(FATAL_ERROR "GGML_CPU_ALL_VARIANTS not yet supported with ${GGML_SYSTEM_ARCH} on ${CMAKE_SYSTEM_NAME}")
	endif()
	elseif (GGML_CPU)
	ggml_add_cpu_backend_variant_impl("")
	endif()

The idea is to avoid adding feature detection logic in the ci/run.sh script and instead consolidate it in the existing CMake implementation for building the possible CPU backends.

[martin-klacer-arm]

Thank you for the comment! That's a very good point, after further investigation I verified that the CMake implementation already does this in a satisfactory way and the existing feature detection/-march setting code from before my patch is actually redundant.

In the latest patch, I removed all the previously existing CPU -march detection, which addresses the same issue as my original patch did and is, I assume, the preferred option.

[ggerganov]

@CISC Do you know if I can start the CI workflow on this branch? Or do I need to clone it in a repo branch to do that?

[CISC]

@CISC Do you know if I can start the CI workflow on this branch? Or do I need to clone it in a repo branch to do that?

I have not found a way to manually trigger workflows on PR-branches, would like to know myself. :)

[ggerganov]

Running here: https://github.com/ggml-org/llama.cpp/actions/runs/23847023744/job/69516801996. If green, we can merge.