feat(qwen35): shard GDR state per TP rank + batched eager/graph decode under TP (+27B benchmark validation) - #946

Open

Ma1oneZhang wants to merge 5 commits into

pegainfer-project:mainfrom

Ma1oneZhang:feat/qwen35-tp2-batched-decode

Commits on Aug 20, 2026

feat(qwen35): shard linear-attention/GDR state per TP rank
Ma1oneZhang
committed
2530dfc
perf(qwen35): batch eager decode rows under TP
Ma1oneZhang
committed
b3aa5e1
docs(qwen35): record the TP rebase onto #870
Ma1oneZhang
committed
b376772
docs(qwen35): TP implementation record + 27B TP2 knowledge benchmark eval
Ma1oneZhang
committed
f56ce51
feat(qwen35): cuda graph decode under TP
Ma1oneZhang
committed
0570313