From 8a1a9b5126126e5228b95fa909d4b08fac65e8b3 Mon Sep 17 00:00:00 2001 From: ynankani Date: Fri, 9 Oct 2026 06:32:52 +0000 Subject: [PATCH] CUDA: pass src1 precision to host MMQ config helpers (#30168) * CUDA: pass src1 precision to host MMQ config helpers Signed-off-by: ynankani * make src1 prec explicit Signed-off-by: ynankani --------- Signed-off-by: ynankani --- ggml/src/ggml-cuda/mmq-load-tiles.cuh | 128 ++++++++--------- ggml/src/ggml-cuda/mmq-vec-dot.cuh | 80 +++++------ ggml/src/ggml-cuda/mmq.cu | 44 +++--- ggml/src/ggml-cuda/mmq.cuh | 196 +++++++++++++------------- 4 files changed, 223 insertions(+), 225 deletions(-) diff --git a/ggml/src/ggml-cuda/mmq-load-tiles.cuh b/ggml/src/ggml-cuda/mmq-load-tiles.cuh index d5654868b..0e59c0347 100644 --- a/ggml/src/ggml-cuda/mmq-load-tiles.cuh +++ b/ggml/src/ggml-cuda/mmq-load-tiles.cuh @@ -7,9 +7,9 @@ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q1_0( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -98,9 +98,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q2_0( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -187,9 +187,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q4_0( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -250,9 +250,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q4_1( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -313,9 +313,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q5_0( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -393,9 +393,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q5_1( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -471,9 +471,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q8_0( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -537,9 +537,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q2_K( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -598,9 +598,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q3_K( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -711,9 +711,9 @@ static __device__ __forceinline__ int unpack_scales_q45_K(const int * scales, co template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q4_K( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -822,9 +822,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q5_K( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -946,9 +946,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_q6_K( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1036,9 +1036,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_iq1_s( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1098,9 +1098,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_iq2_xxs( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1162,9 +1162,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_iq2_xs( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1227,9 +1227,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_iq2_s( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1295,9 +1295,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_iq3_xxs( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1359,9 +1359,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_iq3_s( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1428,9 +1428,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_iq4_xs( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1495,9 +1495,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_iq4_nl( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1564,9 +1564,9 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_mxfp4( const char * __restrict__ x, int * __restrict__ x_tile, const int kbx0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); #if defined(AMD_MFMA_AVAILABLE) || defined(TURING_MMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) int * x_qs = (int *) x_tile; @@ -1670,7 +1670,7 @@ template static __device__ __forceinline_ } } -template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_nvfp4( +template static __device__ __forceinline__ void ggml_cuda_mmq_load_tiles_nvfp4( const char * __restrict__ x, int * __restrict__ x_tile, const int kb0, const int i_max, const int stride) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, prec_src1) / warp_size; diff --git a/ggml/src/ggml-cuda/mmq-vec-dot.cuh b/ggml/src/ggml-cuda/mmq-vec-dot.cuh index b39e4d579..a75dd046e 100644 --- a/ggml/src/ggml-cuda/mmq-vec-dot.cuh +++ b/ggml/src/ggml-cuda/mmq-vec-dot.cuh @@ -10,8 +10,8 @@ using namespace ggml_cuda_mma; template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q4_0_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(GGML_TYPE_Q4_0, I); const int * x_qs = (const int *) x; @@ -60,8 +60,8 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q4_1_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(GGML_TYPE_Q4_1, I); const int * x_qs = (const int *) x; @@ -110,8 +110,8 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(GGML_TYPE_Q8_0, I); const int * x_qs = (const int *) x; @@ -148,8 +148,8 @@ static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma( typedef tile<16, 8, int, input_layout> tile_B; typedef tile<16, 16, int, DATA_LAYOUT_J_MAJOR> tile_C; - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, GGML_PREC_Q8); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -203,8 +203,8 @@ static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma( typedef tile< 8, 8, int> tile_B; typedef tile<16, 8, int> tile_C; - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, GGML_PREC_Q8); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -281,8 +281,8 @@ static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma( template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q8_1_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(GGML_TYPE_Q5_1, I); const int * x_qs = (const int *) x; @@ -318,8 +318,8 @@ template static __device__ __forceinline_ typedef tile<16, 8, int, input_layout> tile_B; typedef tile<16, 16, int, DATA_LAYOUT_J_MAJOR> tile_C; - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, GGML_PREC_Q8); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -368,8 +368,8 @@ template static __device__ __forceinline_ typedef tile< 8, 8, int> tile_B; typedef tile<16, 8, int> tile_C; - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, GGML_PREC_Q8); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -442,8 +442,8 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(type, I); const int * x_qs = (const int *) x; @@ -474,7 +474,7 @@ template static __device__ __forceinline_ } // Used for Q3_K, IQ2_S, and IQ2_XS: -template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_mma( +template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_mma( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { #if defined(AMD_MFMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) constexpr data_layout input_layout = get_input_data_layout(); @@ -483,7 +483,7 @@ template tile_C; constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, prec_src1); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, prec_src1); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -533,7 +533,7 @@ template tile_C; constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, prec_src1); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, prec_src1); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -610,8 +610,8 @@ template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q2_K_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(GGML_TYPE_Q2_K, I); const int * x_qs = (const int *) x; @@ -680,8 +680,8 @@ template static __device__ __forceinline_ typedef tile<16, 4, int, input_layout> tile_B; typedef tile<16, 16, int, DATA_LAYOUT_J_MAJOR> tile_C; - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, GGML_PREC_Q8); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -749,8 +749,8 @@ template static __device__ __forceinline_ typedef tile< 8, 4, int> tile_B; typedef tile<16, 8, int> tile_C; - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, GGML_PREC_Q8); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -870,8 +870,8 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q3_K_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(GGML_TYPE_Q3_K, I); const int * x_qs = (const int *) x; @@ -905,8 +905,8 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q4_K_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(GGML_TYPE_Q4_K, I); const int * x_qs = (const int *) x; @@ -940,8 +940,8 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q5_K_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(GGML_TYPE_Q5_K, I); const int * x_qs = (const int *) x; @@ -975,8 +975,8 @@ template static __device__ __forceinline_ template static __device__ __forceinline__ void ggml_cuda_mmq_vec_dot_q6_K_q8_1_dp4a( const int * __restrict__ x, const int * __restrict__ y, float * __restrict__ sum, const int k00) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, GGML_PREC_Q8) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, GGML_PREC_Q8); constexpr tile_x_sizes txs = mmq_get_dp4a_tile_x_sizes(GGML_TYPE_Q6_K, I); const int * x_qs = (const int *) x; @@ -1015,8 +1015,8 @@ template static __device__ __forceinline_ typedef tile<16, 4, int, input_layout> tile_B; typedef tile<16, 16, int, DATA_LAYOUT_J_MAJOR> tile_C; - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, GGML_PREC_Q8); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -1066,8 +1066,8 @@ template static __device__ __forceinline_ typedef tile< 8, 4, int> tile_B; typedef tile<16, 8, int> tile_C; - constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q8); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, GGML_PREC_Q8); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. y += (threadIdx.y % ntx) * (tile_C::J*MMQ_TILE_Y_K); @@ -1181,7 +1181,7 @@ template static __device__ __forceinline_ typedef tile<16, 8, float> tile_C; constexpr int sram_stride = ggml_cuda_mmq_get_sram_stride(type, J, fallback, GGML_PREC_Q4); - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, GGML_PREC_Q4); constexpr int ntx = rows_per_warp / tile_C::I; constexpr int nfrags = MMQ_TILE_NE_K / tile_A::J; diff --git a/ggml/src/ggml-cuda/mmq.cu b/ggml/src/ggml-cuda/mmq.cu index 027a7e60d..91b89999e 100644 --- a/ggml/src/ggml-cuda/mmq.cu +++ b/ggml/src/ggml-cuda/mmq.cu @@ -8,66 +8,66 @@ static void ggml_cuda_mul_mat_q_switch_type(ggml_backend_cuda_context & ctx, const mmq_args & args, cudaStream_t stream, const ggml_prec prec_src1) { switch (args.type_x) { case GGML_TYPE_Q1_0: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q2_0: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q4_0: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q4_1: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q5_0: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q5_1: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q8_0: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; // ----------------------------------------------------------------------- case GGML_TYPE_Q2_K: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q3_K: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q4_K: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q5_K: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_Q6_K: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; // ----------------------------------------------------------------------- case GGML_TYPE_IQ1_S: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_IQ2_XXS: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_IQ2_XS: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_IQ2_S: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_IQ3_XXS: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_IQ3_S: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_IQ4_XS: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_IQ4_NL: - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; // ----------------------------------------------------------------------- case GGML_TYPE_MXFP4: @@ -76,14 +76,14 @@ static void ggml_cuda_mul_mat_q_switch_type(ggml_backend_cuda_context & ctx, con mul_mat_q_case(ctx, args, stream); break; } - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; case GGML_TYPE_NVFP4: if (prec_src1 == GGML_PREC_Q4) { mul_mat_q_case(ctx, args, stream); break; } - mul_mat_q_case(ctx, args, stream); + mul_mat_q_case(ctx, args, stream); break; default: GGML_ABORT("fatal error"); diff --git a/ggml/src/ggml-cuda/mmq.cuh b/ggml/src/ggml-cuda/mmq.cuh index 198aab4b5..d3e7b75ea 100644 --- a/ggml/src/ggml-cuda/mmq.cuh +++ b/ggml/src/ggml-cuda/mmq.cuh @@ -227,7 +227,7 @@ struct ggml_cuda_mmq_config { #undef CASE -static __host__ ggml_cuda_mmq_config ggml_cuda_mmq_get_config(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1 = GGML_PREC_Q8) { +static __host__ ggml_cuda_mmq_config ggml_cuda_mmq_get_config(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1) { if (GGML_CUDA_CC_IS_AMD(cc)) { if (GGML_CUDA_CC_IS_GCN(cc)) { return ggml_cuda_mmq_get_config_gcn(type, J, fallback); @@ -262,7 +262,7 @@ static __host__ ggml_cuda_mmq_config ggml_cuda_mmq_get_config(const ggml_type ty return ggml_cuda_mmq_get_config_pascal_older(type, J, fallback); } -static constexpr __device__ ggml_cuda_mmq_config ggml_cuda_mmq_get_config(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ ggml_cuda_mmq_config ggml_cuda_mmq_get_config(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { #ifdef GGML_USE_HIP #ifdef GCN return ggml_cuda_mmq_get_config_gcn(type, J, fallback); @@ -295,79 +295,77 @@ static constexpr __device__ ggml_cuda_mmq_config ggml_cuda_mmq_get_config(ggml_t GGML_UNUSED_VARS(type, J, fallback, prec_src1); } -// FIXME all of the host functions are missing prec_src1, this can lead to inconsitent behavior. - -static __host__ int ggml_cuda_mmq_get_type(const ggml_type type, const int J, const bool fallback, const int cc) { - return ggml_cuda_mmq_get_config(type, J, fallback, cc).type; +static __host__ int ggml_cuda_mmq_get_type(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1) { + return ggml_cuda_mmq_get_config(type, J, fallback, cc, prec_src1).type; } -static constexpr __device__ int ggml_cuda_mmq_get_type(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ int ggml_cuda_mmq_get_type(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).type; } -static constexpr __device__ int ggml_cuda_mmq_get_nthreads(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ int ggml_cuda_mmq_get_nthreads(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).nthreads; } -static constexpr __device__ int ggml_cuda_mmq_get_occupancy(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ int ggml_cuda_mmq_get_occupancy(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).occupancy; } -static __host__ int ggml_cuda_mmq_get_I(const ggml_type type, const int J, const bool fallback, const int cc) { - return ggml_cuda_mmq_get_config(type, J, fallback, cc).I; +static __host__ int ggml_cuda_mmq_get_I(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1) { + return ggml_cuda_mmq_get_config(type, J, fallback, cc, prec_src1).I; } -static constexpr __device__ int ggml_cuda_mmq_get_I(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ int ggml_cuda_mmq_get_I(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).I; } -static __host__ int ggml_cuda_mmq_get_J(const ggml_type type, const int J, const bool fallback, const int cc) { - return ggml_cuda_mmq_get_config(type, J, fallback, cc).J; +static __host__ int ggml_cuda_mmq_get_J(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1) { + return ggml_cuda_mmq_get_config(type, J, fallback, cc, prec_src1).J; } -static constexpr __device__ int ggml_cuda_mmq_get_J(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ int ggml_cuda_mmq_get_J(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).J; } -static __host__ ggml_cuda_mmq_sram_layout ggml_cuda_mmq_get_sram_layout(const ggml_type type, const int J, const bool fallback, const int cc) { - return ggml_cuda_mmq_get_config(type, J, fallback, cc).sram_layout; +static __host__ ggml_cuda_mmq_sram_layout ggml_cuda_mmq_get_sram_layout(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1) { + return ggml_cuda_mmq_get_config(type, J, fallback, cc, prec_src1).sram_layout; } -static constexpr __device__ ggml_cuda_mmq_sram_layout ggml_cuda_mmq_get_sram_layout(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ ggml_cuda_mmq_sram_layout ggml_cuda_mmq_get_sram_layout(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).sram_layout; } -static __host__ int ggml_cuda_mmq_get_K_vram(const ggml_type type, const int J, const bool fallback, const int cc) { - return ggml_cuda_mmq_get_config(type, J, fallback, cc).K_vram; +static __host__ int ggml_cuda_mmq_get_K_vram(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1) { + return ggml_cuda_mmq_get_config(type, J, fallback, cc, prec_src1).K_vram; } -static constexpr __device__ int ggml_cuda_mmq_get_K_vram(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ int ggml_cuda_mmq_get_K_vram(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).K_vram; } -static __host__ bool ggml_cuda_mmq_get_stream_k(const ggml_type type, const int J, const bool fallback, const int cc) { - return ggml_cuda_mmq_get_config(type, J, fallback, cc).stream_k; +static __host__ bool ggml_cuda_mmq_get_stream_k(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1) { + return ggml_cuda_mmq_get_config(type, J, fallback, cc, prec_src1).stream_k; } -static constexpr __device__ bool ggml_cuda_mmq_get_stream_k(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ bool ggml_cuda_mmq_get_stream_k(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).stream_k; } -static __host__ int ggml_cuda_mmq_get_fallback(const ggml_type type, const int J, const bool fallback, const int cc) { - return ggml_cuda_mmq_get_config(type, J, fallback, cc).fallback; +static __host__ int ggml_cuda_mmq_get_fallback(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1) { + return ggml_cuda_mmq_get_config(type, J, fallback, cc, prec_src1).fallback; } -static constexpr __device__ int ggml_cuda_mmq_get_fallback(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ int ggml_cuda_mmq_get_fallback(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).fallback; } // --------------------------------------------------------------------------------------------- -static __host__ int ggml_cuda_mmq_get_sram_stride(const ggml_type type, const int J, const bool fallback, const int cc) { - return ggml_cuda_mmq_get_sram_stride(ggml_cuda_mmq_get_sram_layout(type, J, fallback, cc)); +static __host__ int ggml_cuda_mmq_get_sram_stride(const ggml_type type, const int J, const bool fallback, const int cc, const ggml_prec prec_src1) { + return ggml_cuda_mmq_get_sram_stride(ggml_cuda_mmq_get_sram_layout(type, J, fallback, cc, prec_src1)); } -static constexpr __device__ int ggml_cuda_mmq_get_sram_stride(ggml_type type, int J, bool fallback, ggml_prec prec_src1 = GGML_PREC_Q8) { +static constexpr __device__ int ggml_cuda_mmq_get_sram_stride(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { return ggml_cuda_mmq_get_sram_stride(ggml_cuda_mmq_get_sram_layout(type, J, fallback, prec_src1)); } @@ -375,8 +373,8 @@ static __host__ bool ggml_cuda_mmq_needs_fallback(const int64_t nrows_x) { return nrows_x % 128 != 0; } -static constexpr __device__ int ggml_cuda_mmq_get_rows_per_warp(ggml_type type, int J, bool fallback) { - return ggml_cuda_mmq_get_config(type, J, fallback).rows_per_warp(); +static constexpr __device__ int ggml_cuda_mmq_get_rows_per_warp(ggml_type type, int J, bool fallback, ggml_prec prec_src1) { + return ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).rows_per_warp(); } #define MMQ_DP4A_TXS_Q4_0 tile_x_sizes{I*MMQ_TILE_NE_K + I, I*MMQ_TILE_NE_K/QI4_0 + I/QI4_0, 0} @@ -432,12 +430,12 @@ static __host__ int ggml_cuda_mmq_get_nbytes_shared_x(const ggml_cuda_mmq_config #include "mmq-load-tiles.cuh" #include "mmq-vec-dot.cuh" -template static __device__ __forceinline__ void ggml_cuda_mmq_write_back_dp4a( +template static __device__ __forceinline__ void ggml_cuda_mmq_write_back_dp4a( const float * __restrict__ sum, const int32_t * __restrict__ ids_dst, float * __restrict__ dst, const float * __restrict__ y_scale, const int stride, const int i_max, const int j_max) { constexpr int warp_size = ggml_cuda_get_physical_warp_size(); - constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback) / warp_size; - constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback); + constexpr int nwarps = ggml_cuda_mmq_get_nthreads(type, J, fallback, prec_src1) / warp_size; + constexpr int I = ggml_cuda_mmq_get_I(type, J, fallback, prec_src1); const bool y_scale_used = y_scale != nullptr; @@ -471,7 +469,7 @@ template static __device__ __forceinline_ } } -template +template static __device__ __forceinline__ void ggml_cuda_mmq_write_back_mma( const float * __restrict__ sum, const int * __restrict__ ids_dst, float * __restrict__ dst, const float * __restrict__ y_scale, const int stride, const int i_max, const int j_max) { @@ -482,7 +480,7 @@ static __device__ __forceinline__ void ggml_cuda_mmq_write_back_mma( typedef tile<16, 8, int> tile_C; #endif // defined(AMD_MFMA_AVAILABLE) || defined(AMD_WMMA_AVAILABLE) - constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback); + constexpr int rows_per_warp = ggml_cuda_mmq_get_rows_per_warp(type, J, fallback, prec_src1); constexpr int ntx = rows_per_warp/tile_C::I; // Number of x minitiles per warp. const int i0 = (threadIdx.y / ntx) * (ntx*tile_C::I); @@ -536,7 +534,7 @@ struct ggml_cuda_mmq_util_funcs { vdr(vdr), load_tiles(load_tiles), vec_dot(vec_dot), write_back(write_back) {} }; -template +template static constexpr __device__ ggml_cuda_mmq_util_funcs ggml_cuda_mmq_get_util_funcs() { if (!ggml_cuda_mmq_get_config(type, J, fallback, prec_src1).use_mma_data_layout()) { switch (type) { @@ -545,136 +543,136 @@ static constexpr __device__ ggml_cuda_mmq_util_funcs ggml_cuda_mmq_get_util_func VDR_Q1_0_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q1_0, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q2_0: return ggml_cuda_mmq_util_funcs( VDR_Q2_0_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q2_0, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q4_0: return ggml_cuda_mmq_util_funcs( VDR_Q4_0_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q4_0, ggml_cuda_mmq_vec_dot_q4_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q4_1: return ggml_cuda_mmq_util_funcs( VDR_Q4_1_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q4_1, ggml_cuda_mmq_vec_dot_q4_1_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q5_0: return ggml_cuda_mmq_util_funcs( VDR_Q5_0_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q5_0, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q5_1: return ggml_cuda_mmq_util_funcs( VDR_Q5_1_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q5_1, ggml_cuda_mmq_vec_dot_q8_1_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q8_0: return ggml_cuda_mmq_util_funcs( VDR_Q8_0_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q8_0, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); // --------------------------------------------------------------------------------------------- case GGML_TYPE_Q2_K: return ggml_cuda_mmq_util_funcs( VDR_Q2_K_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q2_K, ggml_cuda_mmq_vec_dot_q2_K_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q3_K: return ggml_cuda_mmq_util_funcs( VDR_Q3_K_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q3_K, ggml_cuda_mmq_vec_dot_q3_K_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q4_K: return ggml_cuda_mmq_util_funcs( VDR_Q4_K_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q4_K, ggml_cuda_mmq_vec_dot_q4_K_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q5_K: return ggml_cuda_mmq_util_funcs( VDR_Q5_K_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q5_K, ggml_cuda_mmq_vec_dot_q5_K_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_Q6_K: return ggml_cuda_mmq_util_funcs( VDR_Q6_K_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_q6_K, ggml_cuda_mmq_vec_dot_q6_K_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); // --------------------------------------------------------------------------------------------- case GGML_TYPE_IQ1_S: return ggml_cuda_mmq_util_funcs( VDR_IQ1_S_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_iq1_s, ggml_cuda_mmq_vec_dot_q8_1_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_IQ2_XXS: return ggml_cuda_mmq_util_funcs( VDR_IQ2_XXS_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_iq2_xxs, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_IQ2_XS: return ggml_cuda_mmq_util_funcs( VDR_IQ2_XS_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_iq2_xs, ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_IQ2_S: return ggml_cuda_mmq_util_funcs( VDR_IQ2_S_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_iq2_s, ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_IQ3_XXS: return ggml_cuda_mmq_util_funcs( VDR_IQ3_XXS_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_iq3_xxs, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_IQ3_S: return ggml_cuda_mmq_util_funcs( VDR_IQ3_S_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_iq3_s, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_IQ4_XS: return ggml_cuda_mmq_util_funcs( VDR_IQ4_XS_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_iq4_xs, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_IQ4_NL: return ggml_cuda_mmq_util_funcs( VDR_IQ4_NL_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_iq4_nl, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); // --------------------------------------------------------------------------------------------- case GGML_TYPE_MXFP4: return ggml_cuda_mmq_util_funcs( VDR_MXFP4_Q8_1_MMQ, ggml_cuda_mmq_load_tiles_mxfp4, ggml_cuda_mmq_vec_dot_q8_0_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); case GGML_TYPE_NVFP4: return ggml_cuda_mmq_util_funcs( VDR_NVFP4_Q8_1_MMQ, - ggml_cuda_mmq_load_tiles_nvfp4, + ggml_cuda_mmq_load_tiles_nvfp4, ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_dp4a, - ggml_cuda_mmq_write_back_dp4a); + ggml_cuda_mmq_write_back_dp4a); default: return ggml_cuda_mmq_util_funcs(1, nullptr, nullptr, nullptr); } @@ -690,7 +688,7 @@ static constexpr __device__ ggml_cuda_mmq_util_funcs ggml_cuda_mmq_get_util_func -1, ggml_cuda_mmq_load_tiles_mxfp4_fp4, ggml_cuda_mmq_vec_dot_fp4_fp4_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); } break; case GGML_TYPE_NVFP4: @@ -699,7 +697,7 @@ static constexpr __device__ ggml_cuda_mmq_util_funcs ggml_cuda_mmq_get_util_func -1, ggml_cuda_mmq_load_tiles_nvfp4_nvfp4, ggml_cuda_mmq_vec_dot_fp4_fp4_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); } break; default: @@ -715,164 +713,164 @@ static constexpr __device__ ggml_cuda_mmq_util_funcs ggml_cuda_mmq_get_util_func -1, ggml_cuda_mmq_load_tiles_q1_0, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q2_0: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q2_0, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q4_0: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q4_0, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q4_1: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q4_1, ggml_cuda_mmq_vec_dot_q8_1_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q5_0: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q5_0, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q5_1: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q5_1, ggml_cuda_mmq_vec_dot_q8_1_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q8_0: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q8_0, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); // --------------------------------------------------------------------------------------------- case GGML_TYPE_Q2_K: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q2_K, ggml_cuda_mmq_vec_dot_q2_K_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q3_K: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q3_K, - ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_mma, + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q4_K: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q4_K, ggml_cuda_mmq_vec_dot_q8_1_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q5_K: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q5_K, ggml_cuda_mmq_vec_dot_q8_1_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_Q6_K: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_q6_K, ggml_cuda_mmq_vec_dot_q6_K_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); // --------------------------------------------------------------------------------------------- case GGML_TYPE_IQ1_S: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_iq1_s, ggml_cuda_mmq_vec_dot_q8_1_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_IQ2_XXS: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_iq2_xxs, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_IQ2_XS: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_iq2_xs, - ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_mma, + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_IQ2_S: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_iq2_s, - ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_mma, + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_IQ3_XXS: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_iq3_xxs, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_IQ3_S: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_iq3_s, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_IQ4_XS: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_iq4_xs, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_IQ4_NL: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_iq4_nl, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); // --------------------------------------------------------------------------------------------- case GGML_TYPE_MXFP4: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_mxfp4, ggml_cuda_mmq_vec_dot_q8_0_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); case GGML_TYPE_NVFP4: return ggml_cuda_mmq_util_funcs( -1, ggml_cuda_mmq_load_tiles_nvfp4, ggml_cuda_mmq_vec_dot_q8_0_16_q8_1_mma, - ggml_cuda_mmq_write_back_mma); + ggml_cuda_mmq_write_back_mma); default: return ggml_cuda_mmq_util_funcs(1, nullptr, nullptr, nullptr); } } -template +template static constexpr __device__ int ggml_cuda_mmq_get_vdr() { return ggml_cuda_mmq_get_util_funcs().vdr; } -template +template static constexpr __device__ ggml_cuda_mmq_load_tiles_t ggml_cuda_mmq_get_load_tiles() { return ggml_cuda_mmq_get_util_funcs().load_tiles; } -template +template static constexpr __device__ ggml_cuda_mmq_vec_dot_t ggml_cuda_mmq_get_vec_dot() { return ggml_cuda_mmq_get_util_funcs().vec_dot; } -template +template static constexpr __device__ ggml_cuda_mmq_write_back_t ggml_cuda_mmq_get_write_back() { return ggml_cuda_mmq_get_util_funcs().write_back; } // --------------------------------------------------------------------------------------------- -template +template static __device__ __forceinline__ void mul_mat_q_process_tile( const char * __restrict__ x, const int offset_x, const int * __restrict__ y, const int * __restrict__ ids_dst, float * __restrict__ dst, float * __restrict__ tmp_fixup, @@ -953,7 +951,7 @@ static __device__ __forceinline__ void mul_mat_q_process_tile( // The mul_mat_q kernel implements "stream-k" work partitioning as described in https://arxiv.org/abs/2301.03598 -template +template __launch_bounds__(ggml_cuda_mmq_get_nthreads(type, J, fallback, prec_src1), ggml_cuda_mmq_get_occupancy(type, J, fallback, prec_src1)) static __global__ void mul_mat_q( const char * __restrict__ x, const int * __restrict__ y, const int32_t * __restrict__ ids_dst, @@ -1240,7 +1238,7 @@ static __global__ void mul_mat_q( tile_x_max_i, tile_y_max_j, kb0_start, kb0_stop); } -template +template __launch_bounds__(ggml_cuda_mmq_get_nthreads(type, J, fallback, prec_src1)/2, 1) static __global__ void mul_mat_q_stream_k_fixup( const int32_t * __restrict__ ids_dst, const int32_t * __restrict__ expert_bounds, float * __restrict__ dst, @@ -1395,7 +1393,7 @@ static size_t mmq_get_nbytes_shared(const ggml_cuda_mmq_config & config, const i return nbs_ids + nbs_x + GGML_PAD(nbs_y, config.nthreads*sizeof(int)); } -template +template static void launch_mul_mat_q(ggml_backend_cuda_context & ctx, const mmq_args & args, cudaStream_t stream) { const int id = ggml_cuda_get_device(); const int cc = ggml_cuda_info().devices[id].cc; @@ -1477,7 +1475,7 @@ static void launch_mul_mat_q(ggml_backend_cuda_context & ctx, const mmq_args & a ntx_fd); } -template +template void mul_mat_q_switch_J(ggml_backend_cuda_context & ctx, const mmq_args & args, cudaStream_t stream) { switch (args.J_best) { case 8: @@ -1535,7 +1533,7 @@ void mul_mat_q_switch_J(ggml_backend_cuda_context & ctx, const mmq_args & args, } } -template +template void mul_mat_q_case(ggml_backend_cuda_context & ctx, const mmq_args & args, cudaStream_t stream) { if (ggml_cuda_mmq_needs_fallback(args.nrows_x)) { constexpr bool fallback = true; @@ -1547,7 +1545,7 @@ void mul_mat_q_case(ggml_backend_cuda_context & ctx, const mmq_args & args, cuda } #define DECL_MMQ_CASE(type) \ - template void mul_mat_q_case(ggml_backend_cuda_context & ctx, const mmq_args & args, cudaStream_t stream) \ + template void mul_mat_q_case(ggml_backend_cuda_context & ctx, const mmq_args & args, cudaStream_t stream) \ // FP4 variant: uses native FP4 MMA instead of keeping src1 at Q8_1. #define DECL_MMQ_CASE_W4A4(type) \