flagos-ai / flagos-ai/FlagTree
[Kunlunxin] 昆仑芯XPU 使用flagtree 编译Triton Attention 后端报错了
Open
- Dominant language
- Python
- Stars
- 350
- Forks
- 149
- Avg merge
- 2d 4h
- Merged PRs (30d)
- 81
Description
flagtree 编译 vllm 0.20.0 的 triton attention 后端报错了
复现环境
- 镜像: 使用flagtree xpu 3.6 镜像安装vllm 0.20.0 即可。可以把TRITON_ATTN 对应的 triton kernel 拿出来验证一下。
详细日志如下:
[昆仑芯用flagtree 编译TRITON_ATTN后端报错.txt](https://github.com/user-attachments/files/30774319/flagtree.TRITON_ATTN.txt)
Contributor guide
Research direction
Start by reproducing the failure in the FlagTree XPU 3.6 image with vLLM 0.20.0, then read the attached log and isolate the TRITON_ATTN Triton kernel as suggested. Done means the TRITON_ATTN backend compiles successfully in that environment without the reported error.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- compilers
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100