Skip to content

[feat] enable fused qknorm and rope#20

Merged
qichu-yun merged 4 commits intozejunchen-zejun:dev/perffrom
amd-youchen:dev-rope-qknorm-fused-new
Nov 20, 2025
Merged

[feat] enable fused qknorm and rope#20
qichu-yun merged 4 commits intozejunchen-zejun:dev/perffrom
amd-youchen:dev-rope-qknorm-fused-new

Conversation

@amd-youchen
Copy link

@amd-youchen amd-youchen commented Nov 18, 2025

Can only be merged after

ROCm/aiter#1406
ROCm/aiter#1427

@amd-youchen
Copy link
Author

python3 bench_sglang.py --port 30001 --concurrency 128

image

@amd-youchen
Copy link
Author

Before

image

After

image

@qichu-yun qichu-yun merged commit ace9f2c into zejunchen-zejun:dev/perf Nov 20, 2025
1 check failed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants