Kazeia
/
Kazeia-engine
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

pp1024 plat (saturation batchable), bornes GDN: cout B-2 non mesure, profiler bloque par cli

This commit is contained in:
Richard Loyer 2026-05-24 15:44:45 +02:00
parent 4dd863ecb5
commit 753d8be2b0
1 changed files with 3 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
CLAUDE.md
View File

@ -46,3 +46,6 @@ prefill HTP pp128=44 / decode CPU tg16=13.4 ; HTP decode=4.4, CPU prefill pp32=1
## Verdict 24/05 (mesure honnête, pp égal) ## Verdict 24/05 (mesure honnête, pp égal)
prefill pp32/128/512: CPU 18.5/18.7/18.7 (PLAT), HTP 18.5/44/47. HTP ×2.5 reel a pp512. GDN HTP = decode-only (ggml-hexagon.cpp:2284 "n_tokens!=1 -> CPU fallback"). Prefill GDN sur CPU serie = ancre. Kernel GDN-prefill chunked HTP = le sujet B, gros gain (FA/FFN HTP pourraient ~159). cli boucle = bug runtime aval, PAS compile (bench=meme graphe). Decode tranche, prefill = chantier. prefill pp32/128/512: CPU 18.5/18.7/18.7 (PLAT), HTP 18.5/44/47. HTP ×2.5 reel a pp512. GDN HTP = decode-only (ggml-hexagon.cpp:2284 "n_tokens!=1 -> CPU fallback"). Prefill GDN sur CPU serie = ancre. Kernel GDN-prefill chunked HTP = le sujet B, gros gain (FA/FFN HTP pourraient ~159). cli boucle = bug runtime aval, PAS compile (bench=meme graphe). Decode tranche, prefill = chantier.
## pp1024 + bornes GDN (24/05)
HTP pp512=45/pp1024=45 plat: partie batchable saturee des pp128 (ne prouve PAS la taille du GDN). CPU 18.7=53.5ms/tok modele entier (PAS GDN isole). HTP 22.2ms = GDN_cpu+R_htp, donc GDN_cpu<22.2 borne haute. Recompense B-2 entre ~62 et ~160 t/s selon GDN_cpu non mesure. Profiler HTP (GGML_HEXAGON_PROFILE+OPSYNC) inerte: cli pend AU prefill, bench masque DEBUG. Mesurer GDN_cpu = prerequis avant engager kernel; debloquer cli ouvre le profiler. Existence B chiffree, recompense NON.