CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-03 09:22:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

CUDA: faster dequantize kernels for Q4_0 and Q4_1 (#4938)

Browse Source 
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2024-01-15 07:48:06 +02:00
committed by GitHub
parent a836c8f534
commit 4a3156de2f
1 changed files with 73 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

77
ggml-cuda.cu
View File

@@ -1105,6 +1105,61 @@ static __device__ __forceinline__ void dequantize_q8_0(const void * vx, const in
#endif // GGML_CUDA_F16
}
template<typename dst_t>
static __global__ void dequantize_block_q4_0(const void * __restrict__ vx, dst_t * __restrict__ yy, int nb32) {
const int i = blockIdx.x;
// assume 32 threads
const int tid = threadIdx.x;
const int il = tid/8;
const int ir = tid%8;
const int ib = 8*i + ir;
if (ib >= nb32) {
return;
}
dst_t * y = yy + 256*i + 32*ir + 4*il;
const block_q4_0 * x = (const block_q4_0 *)vx + ib;
const float d = __half2float(x->d);
const float dm = -8*d;
const uint8_t * q = x->qs + 4*il;
for (int l = 0; l < 4; ++l) {
y[l+ 0] = d * (q[l] & 0xF) + dm;
y[l+16] = d * (q[l] >> 4) + dm;
}
}
template<typename dst_t>
static __global__ void dequantize_block_q4_1(const void * __restrict__ vx, dst_t * __restrict__ yy, int nb32) {
const int i = blockIdx.x;
// assume 32 threads
const int tid = threadIdx.x;
const int il = tid/8;
const int ir = tid%8;
const int ib = 8*i + ir;
if (ib >= nb32) {
return;
}
dst_t * y = yy + 256*i + 32*ir + 4*il;
const block_q4_1 * x = (const block_q4_1 *)vx + ib;
const float2 d = __half22float2(x->dm);
const uint8_t * q = x->qs + 4*il;
for (int l = 0; l < 4; ++l) {
y[l+ 0] = d.x * (q[l] & 0xF) + d.y;
y[l+16] = d.x * (q[l] >> 4) + d.y;
}
}
//================================== k-quants
template<typename dst_t>
@@ -6253,6 +6308,20 @@ static void dequantize_row_q3_K_cuda(const void * vx, dst_t * y, const int k, cu
#endif
}
template<typename dst_t>
static void dequantize_q4_0_cuda(const void * vx, dst_t * y, const int k, cudaStream_t stream) {
const int nb32 = k / 32;
const int nb = (k + 255) / 256;
dequantize_block_q4_0<<<nb, 32, 0, stream>>>(vx, y, nb32);
}
template<typename dst_t>
static void dequantize_q4_1_cuda(const void * vx, dst_t * y, const int k, cudaStream_t stream) {
const int nb32 = k / 32;
const int nb = (k + 255) / 256;
dequantize_block_q4_1<<<nb, 32, 0, stream>>>(vx, y, nb32);
}
template<typename dst_t>
static void dequantize_row_q4_K_cuda(const void * vx, dst_t * y, const int k, cudaStream_t stream) {
const int nb = k / QK_K;
@@ -6301,9 +6370,9 @@ static to_fp16_cuda_t ggml_get_to_fp16_cuda(ggml_type type) {
int id;
switch (type) {
case GGML_TYPE_Q4_0:
return dequantize_block_cuda<QK4_0, QR4_0, dequantize_q4_0>;
return dequantize_q4_0_cuda;
case GGML_TYPE_Q4_1:
return dequantize_block_cuda<QK4_1, QR4_1, dequantize_q4_1>;
return dequantize_q4_1_cuda;
case GGML_TYPE_Q5_0:
return dequantize_block_cuda<QK5_0, QR5_0, dequantize_q5_0>;
case GGML_TYPE_Q5_1:
@@ -6338,9 +6407,9 @@ static to_fp16_cuda_t ggml_get_to_fp16_cuda(ggml_type type) {
static to_fp32_cuda_t ggml_get_to_fp32_cuda(ggml_type type) {
switch (type) {
case GGML_TYPE_Q4_0:
return dequantize_block_cuda<QK4_0, QR4_0, dequantize_q4_0>;
return dequantize_q4_0_cuda;
case GGML_TYPE_Q4_1:
return dequantize_block_cuda<QK4_1, QR4_1, dequantize_q4_1>;
return dequantize_q4_1_cuda;
case GGML_TYPE_Q5_0:
return dequantize_block_cuda<QK5_0, QR5_0, dequantize_q5_0>;
case GGML_TYPE_Q5_1: