ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

[SYCL] Fix SYCL im2col and convert Overflow with Large Dims (#9052)

Browse source 
* sycl: fix im2col overflow and sync with cuda

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>

* sycl: fix convert overflow

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>

* sycl: fix convert and dequantize

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>

* sycl: fix ib in dmmv

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>

* sycl:refine convert

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>

* sycl: move downsample global_range into common

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>

* test: add im2col and convert test cases

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>

* test: make new cases only in sycl

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>

* test: comment new test_cases for only local testing

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>

---------

Signed-off-by: zhentaoyu <zhentao.yu@intel.com>
This commit is contained in:
zhentaoyu 2024-08-20 23:06:51 +08:00 • committed by GitHub
parent 90db8146d5
commit 4f8d19ff17
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
11 changed files with 333 additions and 262 deletions
Download patch file Download diff file Expand all files Collapse all files

4
ggml/src/ggml-sycl/dmmv.cpp
View file

@ -4,7 +4,7 @@
#include "presets.hpp"
static void convert_f16(const void * vx, const int ib, const int iqs, dfloat2 & v){
static void convert_f16(const void * vx, const int64_t ib, const int iqs, dfloat2 & v){
const sycl::half *x = (const sycl::half *)vx;
// automatic half -> float type cast if dfloat == float
@ -12,7 +12,7 @@ static void convert_f16(const void * vx, const int ib, const int iqs, dfloat2 &
v.y() = x[ib + iqs + 1];
}
static void convert_f32(const void * vx, const int ib, const int iqs, dfloat2 & v){
static void convert_f32(const void * vx, const int64_t ib, const int iqs, dfloat2 & v){
const float * x = (const float *) vx;
// automatic half -> float type cast if dfloat == float