root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 03:14:35 +00:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

CLBlast: Fix temporary buffer size for f16 conversion (wsize)

Browse Source 
Fix buffer overflow.
Reduce the size to fit just one 2D slice.
Assert sufficient size.
This commit is contained in:
shibe2 2023-10-11 21:30:06 +04:00
parent a5e8c1d8c7
commit 40e5ce054f
1 changed files with 7 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
ggml-opencl.cpp
View File

@ -1568,7 +1568,7 @@ static void ggml_cl_mul_mat_f32(const ggml_tensor * src0, const ggml_tensor * sr
ggml_cl_pool_free(d_D, d_size); ggml_cl_pool_free(d_D, d_size);
} }
static void ggml_cl_mul_mat_f16(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, void * wdata, size_t /* wsize */) { static void ggml_cl_mul_mat_f16(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, void * wdata, size_t wsize) {
GGML_ASSERT(fp16_support); GGML_ASSERT(fp16_support);
const int64_t ne00 = src0->ne[0]; const int64_t ne00 = src0->ne[0];
@ -1598,6 +1598,10 @@ static void ggml_cl_mul_mat_f16(const ggml_tensor * src0, const ggml_tensor * sr
const int y_ne = ne11 * ne10; const int y_ne = ne11 * ne10;
const int d_ne = ne11 * ne01; const int d_ne = ne11 * ne01;
GGML_ASSERT(wsize >= sizeof(ggml_fp16_t) * y_ne);
GGML_ASSERT(wsize >= sizeof(ggml_fp16_t) * d_ne);
ggml_fp16_t * const tmp = (ggml_fp16_t *) wdata;
size_t x_size; size_t x_size;
size_t y_size; size_t y_size;
size_t d_size; size_t d_size;
@ -1634,7 +1638,6 @@ static void ggml_cl_mul_mat_f16(const ggml_tensor * src0, const ggml_tensor * sr
// convert src1 to fp16 // convert src1 to fp16
// TODO: use multiple threads // TODO: use multiple threads
ggml_fp16_t * const tmp = (ggml_fp16_t *) wdata + (ne11 * ne10) * (i13 * ne12 + i12);
char * src1i = (char *) src1->data + i13*nb13 + i12*nb12; char * src1i = (char *) src1->data + i13*nb13 + i12*nb12;
if (src1_cont_rows) { if (src1_cont_rows) {
if (src1_cont_cols) { if (src1_cont_cols) {
@ -1897,8 +1900,8 @@ void ggml_cl_mul_mat(const struct ggml_tensor * src0, const struct ggml_tensor *
} }
size_t ggml_cl_mul_mat_get_wsize(const struct ggml_tensor * src0, const struct ggml_tensor * src1, struct ggml_tensor * dst) { size_t ggml_cl_mul_mat_get_wsize(const struct ggml_tensor * src0, const struct ggml_tensor * src1, struct ggml_tensor * dst) {
if (ggml_cl_mul_mat_use_f16(src0, src1, dst)) { if (src0->type == GGML_TYPE_F16 && ggml_cl_mul_mat_use_f16(src0, src1, dst)) {
return ggml_nelements(src1) * sizeof(ggml_fp16_t); return sizeof(ggml_fp16_t) * std::max(src1->ne[0] * src1->ne[1], dst->ne[0] * dst->ne[1]);
} }
return 0; return 0;
} }