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

update news

Browse source
This commit is contained in:
Jianyu Zhang 2024-03-01 15:52:41 +08:00
parent 4c29df303d
commit 47a572df16
1 changed files with 9 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

13
README-sycl.md
View file

@ -29,8 +29,13 @@ For Intel CPU, recommend to use llama.cpp for X86 (Intel MKL building).
## News ## News
- 2024.3 - 2024.3
- Support multiple cards: **--split-mode**: none|layer|row - Support multiple cards: **--split-mode**: [none|layer]; not support [row], it's on developing.
- Support to assign main GPU by **--main-gpu** - Support to assign main GPU by **--main-gpu**, replace $GGML_SYCL_DEVICE.
- Support detecting all GPUs with level-zero and same top **Max compute units**.
- Support OPs
- hardsigmoid
- hardswish
- pool2d
- 2024.1 - 2024.1
- Create SYCL backend for Intel GPU. - Create SYCL backend for Intel GPU.
@ -470,9 +475,9 @@ Using device **0** (Intel(R) Arc(TM) A770 Graphics) as main device
Solution: add **--no-mmap** or **--mmap 0**. Solution: add **--no-mmap** or **--mmap 0**.
- Run with split-mode: row is slow - Split-mode: [row] is not supported
It's limited by the missed function: memory copy from device to device. We use dirty solution. It will be fixed in the future. It's on developing.
## Q&A ## Q&A