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

fix cuda pad/scale/im2col3d

Browse Source
This commit is contained in:
leejet
2025-08-12 23:44:46 +08:00
parent 85c8e1e519
commit ae47caca70
3 changed files with 20 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
ggml/src/ggml-cuda/im2col.cu
View File

@@ -142,9 +142,9 @@ static __global__ void im2col_3d_kernel(
const int64_t iih = ioh * s1 + ikh * d1 - p1;
const int64_t iid = iod * s2 + ikd * d2 - p2;
const int64_t offset_dst = (in*OH_OW + ioh*OW + iow)*IC_KD_KH_KW + iic*KD_KH_KW + ikd * KH_KW + ikh*KW + ikw;
const int64_t offset_dst = (in*OD*OH_OW + iod*OH_OW + ioh*OW + iow)*IC_KD_KH_KW + iic*KD_KH_KW + ikd * KH_KW + ikh*KW + ikw;
if (iih < 0 || iih >= IH || iiw < 0 || iiw >= IW) {
if (iih < 0 || iih >= IH || iiw < 0 || iiw >= IW || iid < 0 || iid >= ID) {
dst[offset_dst] = 0.0f;
} else {
const int64_t offset_src = (in*IC + iic)*ID_IH_IW + iid*IH_IW + iih*IW + iiw;

14
ggml/src/ggml-cuda/pad.cu
View File

@@ -22,13 +22,13 @@ static __global__ void pad_f32(const float * src, float * dst,
&& (i1 >= lp1 && i1 < ne1 - rp1) \
&& (i2 >= lp2 && i2 < ne2 - rp2) \
&& (i3 >= lp3 && i3 < ne3 - rp3)) {
int i00 = i0 - lp0;
int i01 = i1 - lp1;
int i02 = i2 - lp2;
int i03 = i3 - lp3;
int ne02 = ne2 - lp2 - rp2;
int ne01 = ne1 - lp1 - rp1;
int ne00 = ne0 - lp0 - rp0;
int64_t i00 = i0 - lp0;
int64_t i01 = i1 - lp1;
int64_t i02 = i2 - lp2;
int64_t i03 = i3 - lp3;
int64_t ne02 = ne2 - lp2 - rp2;
int64_t ne01 = ne1 - lp1 - rp1;
int64_t ne00 = ne0 - lp0 - rp0;
int64_t src_idx = i03*(ne00*ne01*ne02) + i02*(ne00*ne01) + i01*ne00 + i00;

20
ggml/src/ggml-cuda/scale.cu
View File

@@ -1,18 +1,20 @@
#include "scale.cuh"
static __global__ void scale_f32(const float * x, float * dst, const float scale, const float bias, const int k) {
const int i = blockDim.x*blockIdx.x + threadIdx.x;
#define MIN(a, b) (a) < (b) ? (a) : (b)
#define MAX_GRIDDIM_X 0x7FFFFFFF
if (i >= k) {
return;
static __global__ void scale_f32(const float * x, float * dst, const float scale, const float bias, const int64_t nelements) {
int64_t tid = (int64_t)blockIdx.x * (int64_t)blockDim.x + (int64_t)threadIdx.x;
int64_t stride = (int64_t)blockDim.x * (int64_t)gridDim.x;
for (int64_t i = tid; i < nelements; i += stride) {
dst[i] = scale * x[i] + bias;
}
dst[i] = scale * x[i] + bias;
}
static void scale_f32_cuda(const float * x, float * dst, const float scale, const float bias, const int k, cudaStream_t stream) {
const int num_blocks = (k + CUDA_SCALE_BLOCK_SIZE - 1) / CUDA_SCALE_BLOCK_SIZE;
scale_f32<<<num_blocks, CUDA_SCALE_BLOCK_SIZE, 0, stream>>>(x, dst, scale, bias, k);
static void scale_f32_cuda(const float * x, float * dst, const float scale, const float bias, const int64_t nelements, cudaStream_t stream) {
const int64_t num_blocks = (nelements + CUDA_SCALE_BLOCK_SIZE - 1) / CUDA_SCALE_BLOCK_SIZE;
scale_f32<<<MIN(MAX_GRIDDIM_X, num_blocks), CUDA_SCALE_BLOCK_SIZE, 0, stream>>>(x, dst, scale, bias, nelements);
}
void ggml_cuda_op_scale(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {