From 77b6c819dce9bb71b10badba138effabaf572577 Mon Sep 17 00:00:00 2001 From: Jens Elofsson Date: Wed, 29 Jan 2025 13:35:26 +0100 Subject: [PATCH] Remove KAI_ASM_END that causes issues on MSVC Signed-off-by: Jens Elofsson --- ...mul_clamp_f32_qai8dxp4x4_qsi4c32p4x4_16x4_neon_dotprod_asm.S | 2 -- ...mul_clamp_f32_qai8dxp4x8_qsi4c32p4x8_16x4x32_neon_i8mm_asm.S | 2 -- 2 files changed, 4 deletions(-) diff --git a/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi4c32p/kai_matmul_clamp_f32_qai8dxp4x4_qsi4c32p4x4_16x4_neon_dotprod_asm.S b/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi4c32p/kai_matmul_clamp_f32_qai8dxp4x4_qsi4c32p4x4_16x4_neon_dotprod_asm.S index 1a448c5f..f4bc5a6f 100644 --- a/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi4c32p/kai_matmul_clamp_f32_qai8dxp4x4_qsi4c32p4x4_16x4_neon_dotprod_asm.S +++ b/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi4c32p/kai_matmul_clamp_f32_qai8dxp4x4_qsi4c32p4x4_16x4_neon_dotprod_asm.S @@ -801,8 +801,6 @@ KAI_ASM_LABEL(label_25) // Row tail: Row loop skip ret KAI_ASM_FUNCTION_END(kai_kernel_matmul_clamp_f32_qai8dxp4x4_qsi4c32p4x4_16x4_neon_dotprod) - KAI_ASM_END - // Optimized kernel for bl = 32 KAI_ASM_CODE(matmul_clamp_f32_qai8dxp4x4_qsi4c32p4x4_16x4_opt32_neon_dotprod) diff --git a/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi4c32p/kai_matmul_clamp_f32_qai8dxp4x8_qsi4c32p4x8_16x4x32_neon_i8mm_asm.S b/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi4c32p/kai_matmul_clamp_f32_qai8dxp4x8_qsi4c32p4x8_16x4x32_neon_i8mm_asm.S index cc3c75bc..52aa4801 100644 --- a/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi4c32p/kai_matmul_clamp_f32_qai8dxp4x8_qsi4c32p4x8_16x4x32_neon_i8mm_asm.S +++ b/kai/ukernels/matmul/matmul_clamp_f32_qai8dxp_qsi4c32p/kai_matmul_clamp_f32_qai8dxp4x8_qsi4c32p4x8_16x4x32_neon_i8mm_asm.S @@ -741,8 +741,6 @@ KAI_ASM_LABEL(label_25) // Row tail: Row loop skip ret KAI_ASM_FUNCTION_END(kai_kernel_matmul_clamp_f32_qai8dxp4x8_qsi4c32p4x8_16x4x32_neon_i8mm) - KAI_ASM_END - // Optimized kernel for bl = 32 KAI_ASM_CODE(matmul_clamp_f32_qai8dxp4x8_qsi4c32p4x8_16x4x32_opt32_neon_i8mm) -- GitLab