CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-13 10:57:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

[SYCL] Optimize mul_mat for Q4_0 on Intel GPU (#12035)

Browse Source 
* opt performance by reorder for Intel GPU

* detect hw type and save opt feature, and print opt feature

* correct name

* support optimize graph once when compute graph, record the opt status in tensor->extra, make CI passed

* add env variable GGML_SYCL_DISABLE_OPT for debug

* use syclex::architecture replace the custom hw define, update the guide for GGML_SYCL_DISABLE_OPT

* add performance data

* mv getrows functions to separeted files

* fix global variables

---------

Co-authored-by: arthw <14088817+arthw@users.noreply.github.com>
This commit is contained in:
Neo Zhang Jianyu
2025-02-24 22:33:23 +08:00
committed by GitHub
parent 651adf4b66
commit 08d5986290
14 changed files with 803 additions and 266 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
ggml/src/ggml-sycl/common.cpp
View File

@@ -99,3 +99,20 @@ catch (sycl::exception const &exc) {
<< ", line:" << __LINE__ << std::endl;
std::exit(1);
}
void release_extra_gpu(ggml_tensor_extra_gpu * extra, std::vector<queue_ptr> streams) {
for (int i = 0; i < ggml_sycl_info().device_count; ++i) {
for (int64_t is = 0; is < GGML_SYCL_MAX_STREAMS; ++is) {
if (extra->events[i][is] != nullptr) {
SYCL_CHECK(CHECK_TRY_ERROR(dpct::destroy_event(extra->events[i][is])));
}
}
if (extra->data_device[i] != nullptr && streams.size()>0) {
ggml_sycl_set_device(i);
SYCL_CHECK(
CHECK_TRY_ERROR(sycl::free(extra->data_device[i], *(streams[i]))));
}
}
delete extra;
}