[None][fix] Fix disagg KV cache transfer crash on reuse budget skip - #12918

Closed

reasonsolo wants to merge 14 commits into

NVIDIA:mainfrom

reasonsolo:feat/bench_y

Commits on Mar 31, 2026

[None][perf] Split MLA DSA custom op for piecewise CUDA graph capture
liji-nv
committed
acefe29
[None][fix] Fix pre_indexer bogus parallel pattern and mla_dsa_proj docstring
liji-nv
committed
c8466f0
[None][refactor] Move _update_k_cache into sparse_attn_indexer
liji-nv
committed
f0dc05f
[None][chore] Clean up MLA DSA custom op dispatch
liji-nv
committed
66627a1
[None][chore] Restore quant_block_size assertion in sparse_attn_indexer
liji-nv
committed
c21c41c

Commits on Apr 2, 2026

Commits on Apr 3, 2026

[https://nvbugs/5983390][perf] Reduce host overhead in DSA MLA attention path (#12691)
liji-nv
authored
ca4ace1

Commits on Apr 6, 2026

Add default values for num_contexts and num_ctx_tokens
v-shobhit
authored
4140ccc

Commits on Apr 7, 2026

[None][fix] Pre-subtract non-first-chunk context costs in reuse budge… (#12806)
liji-nv
authored
01c4947

Commits on Apr 10, 2026