From 8b7132972db277e1841d928d2601716545f8429a Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Fri, 19 May 2023 19:00:44 +0300 Subject: [PATCH] cuda : update Q4 and Q8 dequantize kernels --- ggml-cuda.cu | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/ggml-cuda.cu b/ggml-cuda.cu index f2630ec8e..688bcf799 100644 --- a/ggml-cuda.cu +++ b/ggml-cuda.cu @@ -42,19 +42,19 @@ typedef void (*dequantize_mul_mat_vec_cuda_t)(const void * vx, const float * y, #define QK4_0 32 #define QR4_0 2 typedef struct { - float d; // delta + half d; // delta uint8_t qs[QK4_0 / 2]; // nibbles / quants } block_q4_0; -static_assert(sizeof(block_q4_0) == sizeof(float) + QK4_0 / 2, "wrong q4_0 block size/padding"); +static_assert(sizeof(block_q4_0) == sizeof(ggml_fp16_t) + QK4_0 / 2, "wrong q4_0 block size/padding"); #define QK4_1 32 #define QR4_1 2 typedef struct { - float d; // delta - float m; // min + half d; // delta + half m; // min uint8_t qs[QK4_1 / 2]; // nibbles / quants } block_q4_1; -static_assert(sizeof(block_q4_1) == sizeof(float) * 2 + QK4_1 / 2, "wrong q4_1 block size/padding"); +static_assert(sizeof(block_q4_1) == sizeof(ggml_fp16_t) * 2 + QK4_1 / 2, "wrong q4_1 block size/padding"); #define QK5_0 32 #define QR5_0 2 @@ -78,10 +78,10 @@ static_assert(sizeof(block_q5_1) == 2 * sizeof(ggml_fp16_t) + sizeof(uint32_t) + #define QK8_0 32 #define QR8_0 1 typedef struct { - float d; // delta + half d; // delta int8_t qs[QK8_0]; // quants } block_q8_0; -static_assert(sizeof(block_q8_0) == sizeof(float) + QK8_0, "wrong q8_0 block size/padding"); +static_assert(sizeof(block_q8_0) == sizeof(ggml_fp16_t) + QK8_0, "wrong q8_0 block size/padding"); #define CUDA_DEQUANTIZE_BLOCK_SIZE 256 #define CUDA_DMMV_BLOCK_SIZE 32 // dmmv = dequantize_mul_mat_vec