]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
[SYCL] refactor (#6408)
authorMeng, Hengyu <redacted>
Wed, 19 Jun 2024 01:11:51 +0000 (09:11 +0800)
committerGitHub <redacted>
Wed, 19 Jun 2024 01:11:51 +0000 (09:11 +0800)
commit623494a478134432fd2d7ee40135770a3340674f
tree234f90caed1bbe0601cd4f6239e0f2885c34bb6a
parent37bef8943312d91183ff06d8f1214082a17344a5
[SYCL] refactor (#6408)

* seperate lower precision GEMM from the main files

* fix workgroup size hardcode
13 files changed:
ggml-sycl.cpp
ggml-sycl/backend.hpp
ggml-sycl/convert.cpp [new file with mode: 0644]
ggml-sycl/convert.hpp [new file with mode: 0644]
ggml-sycl/dequantize.hpp [new file with mode: 0644]
ggml-sycl/dmmv.cpp [new file with mode: 0644]
ggml-sycl/dmmv.hpp [new file with mode: 0644]
ggml-sycl/mmq.cpp [new file with mode: 0644]
ggml-sycl/mmq.hpp [new file with mode: 0644]
ggml-sycl/mmvq.cpp [new file with mode: 0644]
ggml-sycl/mmvq.hpp [new file with mode: 0644]
ggml-sycl/presets.hpp
ggml-sycl/vecdotq.hpp [new file with mode: 0644]