llama.cpp/ggml-cuda/dmmv.cuh

19 lines
584 B
Plaintext
Raw Permalink Normal View History

#include "common.cuh"
// dmmv = dequantize_mul_mat_vec
// TODO: remove this?
#ifndef GGML_CUDA_DMMV_X
#define GGML_CUDA_DMMV_X 32
#endif
#ifndef GGML_CUDA_MMV_Y
#define GGML_CUDA_MMV_Y 1
#endif
void ggml_cuda_op_dequantize_mul_mat_vec(
ggml_backend_cuda_context & ctx,
const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, const char * src0_dd_i, const float * src1_ddf_i,
const char * src1_ddq_i, float * dst_dd_i, const int64_t row_low, const int64_t row_high, const int64_t src1_ncols,
const int64_t src1_padded_row_size, cudaStream_t stream);