From b34e02348064c2f0cef1f89b44d9bee4eb15b9e7 Mon Sep 17 00:00:00 2001 From: R0CKSTAR Date: Wed, 11 Sep 2024 09:46:55 +0800 Subject: [PATCH] musa: remove Clang builtins mapping (#9421) Signed-off-by: Xiaodong Ye --- ggml/src/ggml-cuda/vendors/musa.h | 39 ------------------------------- 1 file changed, 39 deletions(-) diff --git a/ggml/src/ggml-cuda/vendors/musa.h b/ggml/src/ggml-cuda/vendors/musa.h index e50a103ac..8df571149 100644 --- a/ggml/src/ggml-cuda/vendors/musa.h +++ b/ggml/src/ggml-cuda/vendors/musa.h @@ -130,42 +130,3 @@ #define cudaKernelNodeParams musaKernelNodeParams #define cudaStreamCaptureModeRelaxed musaStreamCaptureModeRelaxed #define cudaStreamEndCapture musaStreamEndCapture - -// XXX: Clang builtins mapping -#define __vsub4 __vsub4_musa -#define __vcmpeq4 __vcmpeq4_musa -#define __vcmpne4 __vcmpne4_musa - -#ifndef __has_builtin - #define __has_builtin(x) 0 -#endif - -typedef uint8_t uint8x4_t __attribute__((ext_vector_type(4))); - -static __device__ __forceinline__ int __vsub4_musa(const int a, const int b) { - return __vsubss4(a, b); -} - -static __device__ __forceinline__ unsigned int __vcmpeq4_musa(unsigned int a, unsigned int b) { - const uint8x4_t& va = reinterpret_cast(a); - const uint8x4_t& vb = reinterpret_cast(b); - unsigned int c; - uint8x4_t& vc = reinterpret_cast(c); -#pragma unroll - for (int i = 0; i < 4; ++i) { - vc[i] = va[i] == vb[i] ? 0xff : 0x00; - } - return c; -} - -static __device__ __forceinline__ unsigned int __vcmpne4_musa(unsigned int a, unsigned int b) { - const uint8x4_t& va = reinterpret_cast(a); - const uint8x4_t& vb = reinterpret_cast(b); - unsigned int c; - uint8x4_t& vc = reinterpret_cast(c); -#pragma unroll - for (int i = 0; i < 4; ++i) { - vc[i] = va[i] == vb[i] ? 0x00 : 0xff; - } - return c; -}