Kazeia
/
Kazeia-engine
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
Kazeia-engine/docs/02_cont_overhead.md

455 B
Raw Blame History

Chantier 2 — CONT/copies prefill: ROI faible, non engage

416 CONT = ~17/couche GDN (reshape/permute DeltaNet, qwen3next.cpp). 4.6s mesure OPSYNC=serialise; async (defaut) les copies recouvrent le compute -> part reelle << gonfle. cont l.165 = TODO upstream (CUDA non-contig unary), pas removable safe. prefill deja 46-50, pas de bottleneck CONT prouve. Decision: low-ROI, skip. Le vrai gain prefill = GDN sur HTP (chantier 3), pas tailler des copies.