eusonice 3da116d40d adr-0070: gqa short context attention prefill/decode + 3 composite tiers

2026-06-26 15:59:21 -07:00

..

ADR-0028-par-dtensor-support.md

ADR housekeeping: category prefixes, lifecycle folders, retroactive 0034-0037

2026-05-20 01:15:55 -07:00

ADR-0030-mem-ipcq-physaddr.md

ADR housekeeping: category prefixes, lifecycle folders, retroactive 0034-0037

2026-05-20 01:15:55 -07:00

ADR-0061-prog-tl-broadcast.md

gqa: ADR-0060/0062/0063/0064 unified GQA kernels + CPU cost model

2026-06-09 18:15:59 -07:00

ADR-0070-algo-gqa-short-context-attention.md

adr-0070: gqa short context attention prefill/decode + 3 composite tiers

2026-06-26 15:59:21 -07:00

DDD-0060-gqa-fused-attention-detailed-design.md

gqa(adr): review fixes — unify KV placement to contiguous, ring/reduce cost model, opt3 wording, rank def

2026-06-05 00:52:38 -07:00