Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

sycl : support to malloc memory on device more than 4GB, update the doc and script (#17566)

Browse Source 
Co-authored-by: Neo Zhang Jianyu <jianyu.zhang@intel.com>
This commit is contained in:
Neo Zhang
2025-11-29 14:59:44 +02:00
committed by GitHub
co-authored by Neo Zhang Jianyu
parent f698a79c63
commit 7d2add51d8
7 changed files with 33 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-sycl/CMakeLists.txt
+6 -2
View File
@@ -91,7 +91,10 @@ if (GGML_SYCL_F16)
add_compile_definitions(GGML_SYCL_F16)
endif()
if (GGML_SYCL_TARGET STREQUAL "NVIDIA")
if (GGML_SYCL_TARGET STREQUAL "INTEL")
add_compile_definitions(GGML_SYCL_WARP_SIZE=16)
target_link_options(ggml-sycl PRIVATE -Xs -ze-intel-greater-than-4GB-buffer-required)
elseif (GGML_SYCL_TARGET STREQUAL "NVIDIA")
add_compile_definitions(GGML_SYCL_WARP_SIZE=32)
elseif (GGML_SYCL_TARGET STREQUAL "AMD")
# INFO: Allowed Sub_group_sizes are not consistent through all
@@ -100,7 +103,8 @@ elseif (GGML_SYCL_TARGET STREQUAL "AMD")
# Target archs tested working: gfx1030, gfx1031, (Only tested sub_group_size = 32)
add_compile_definitions(GGML_SYCL_WARP_SIZE=32)
else()
add_compile_definitions(GGML_SYCL_WARP_SIZE=16)
# default for other target
add_compile_definitions(GGML_SYCL_WARP_SIZE=32)
endif()
if (GGML_SYCL_GRAPH)