From 641561058bc1c366530871aa1fd1f7deb4be8fdb Mon Sep 17 00:00:00 2001 From: Henri Vasserman Date: Sat, 12 Aug 2023 10:51:46 +0300 Subject: [PATCH] gfx1100 support --------- Co-authored-by: ardfork <134447697+ardfork@users.noreply.github.com> Co-authored-by: jammm <2500920+jammm@users.noreply.github.com> Co-authored-by: jdecourval <7315817+jdecourval@users.noreply.github.com> --- ggml-cuda.cu | 2 ++ 1 file changed, 2 insertions(+) diff --git a/ggml-cuda.cu b/ggml-cuda.cu index aa4a2e919..1b4755d22 100644 --- a/ggml-cuda.cu +++ b/ggml-cuda.cu @@ -92,6 +92,8 @@ static __device__ __forceinline__ int __vsubss4(const int a, const int b) { static __device__ __forceinline__ int __dp4a(const int a, const int b, int c) { #if defined(__gfx906__) || defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx1030__) c = __builtin_amdgcn_sdot4(a, b, c, false); +#elif defined(__gfx1100__) + c = __builtin_amdgcn_sudot4( true, a, true, b, c, false); #elif defined(__gfx1010__) || defined(__gfx900__) int tmp1; int tmp2;