kleidiai : add sme fp16 compute path for q4_0 gemm on aarch64 (#20043)
This commit is contained in:
@@ -566,9 +566,9 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
||||
|
||||
# Fetch KleidiAI sources:
|
||||
include(FetchContent)
|
||||
set(KLEIDIAI_COMMIT_TAG "v1.16.0")
|
||||
set(KLEIDIAI_COMMIT_TAG "v1.22.0")
|
||||
set(KLEIDIAI_DOWNLOAD_URL "https://github.com/ARM-software/kleidiai/archive/refs/tags/${KLEIDIAI_COMMIT_TAG}.tar.gz")
|
||||
set(KLEIDIAI_ARCHIVE_MD5 "0a9e9008adb6031f9e8cf70dff4a3321")
|
||||
set(KLEIDIAI_ARCHIVE_MD5 "54049037570ab0ee0a0d126b2ba5ece1")
|
||||
|
||||
if (POLICY CMP0135)
|
||||
cmake_policy(SET CMP0135 NEW)
|
||||
@@ -608,6 +608,7 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_qsi8d32p_qsi4c32p/
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi8cxp/
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_fp32_bf16p_bf16p/
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_f16p_qsi4c32p/
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/pack/)
|
||||
|
||||
set(ARCH_FLAGS_TEMP "${ARCH_FLAGS}")
|
||||
@@ -648,7 +649,6 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
||||
|
||||
if (NOT SME_ENABLED MATCHES -1)
|
||||
list(APPEND GGML_KLEIDIAI_SOURCES
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_qsi8d32p_qsi4c32p/kai_matmul_clamp_f32_qsi8d32p1vlx4_qsi4c32p4vlx4_1vlx4vl_sme2_mopa.c
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_qsi8d32p_qsi4c32p/kai_matmul_clamp_f32_qsi8d32p1x4_qsi4c32p4vlx4_1x4vl_sme2_sdot.c
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi8cxp/kai_matmul_clamp_f32_qai8dxp1vlx4_qsi8cxp4vlx4_1vlx4vl_sme2_mopa.c
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi8cxp/kai_matmul_clamp_f32_qai8dxp1vlx4_qsi8cxp4vlx4_1vlx4vl_sme2_mopa_asm.S
|
||||
@@ -656,10 +656,13 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi8cxp/kai_matmul_clamp_f32_qai8dxp1x4_qsi8cxp4vlx4_1x4vl_sme2_dot_asm.S
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_fp32_bf16p_bf16p/kai_matmul_clamp_f32_bf16p2vlx2_bf16p2vlx2_2vlx2vl_sme2_mopa.c
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_fp32_bf16p_bf16p/kai_matmul_clamp_f32_bf16p2vlx2_bf16p2vlx2_2vlx2vl_sme2_mopa_asm.S
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_f16p_qsi4c32p/kai_matmul_clamp_f32_f16p1vlx2_qsi4c32p4vlx2_1vlx4vl_sme2_mopa.c
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/matmul_clamp_f32_f16p_qsi4c32p/kai_matmul_clamp_f32_f16p1vlx2_qsi4c32p4vlx2_1vlx4vl_sme2_mopa_asm.S
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/pack/kai_lhs_pack_bf16p2vlx2_f32_sme.c
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/pack/kai_rhs_pack_kxn_bf16p2vlx2b_f32_x32_sme.c
|
||||
${KLEIDIAI_SRC}/kai/ukernels/matmul/pack/kai_lhs_pack_f16pmrx2_f32_neon.c
|
||||
${KLEIDIAI_SRC}/kai/kai_common_sme_asm.S)
|
||||
set(PRIVATE_ARCH_FLAGS "-fno-tree-vectorize;${PRIVATE_ARCH_FLAGS}+sve+sve2")
|
||||
set(PRIVATE_ARCH_FLAGS "-fno-tree-vectorize;${PRIVATE_ARCH_FLAGS}+sve+sve2+sme2+fp16")
|
||||
endif()
|
||||
|
||||
if (NOT SVE_ENABLED MATCHES -1)
|
||||
|
||||
Reference in New Issue
Block a user