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

GDN chunkwise P0: oracle test-backend-ops 28/28 HTP0 (gate régression), baseline pp512=93, boucle build live; P1 chunk-f32 en cours

This commit is contained in:
Richard Loyer 2026-05-26 16:04:58 +02:00
parent de9543de2a
commit aa363e652f
1 changed files with 2 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
docs/04_gdn_chunkwise.md
View File

@ -14,3 +14,5 @@ Chunk C=64: intra-chunk = matmuls C×S triangulaire, inter-chunk = state. Les ra
2. inner GEMM → HMX fp16, comparer dérive vs CPU (prefill tolère, KV only). Bench 93→? 2. inner GEMM → HMX fp16, comparer dérive vs CPU (prefill tolère, KV only). Bench 93→?
3. VTCM pour state inter-chunk (déjà au tg). Cible 150. 3. VTCM pour state inter-chunk (déjà au tg). Cible 150.
Risque: HMX fp16 accum non bit-exact — OK prefill (decode CPU exact). Frontière, aide partielle structure/debug. Risque: HMX fp16 accum non bit-exact — OK prefill (decode CPU exact). Frontière, aide partielle structure/debug.
## P0 fait (26/05): oracle = `test-backend-ops -o GATED_DELTA_NET -b HTP0` = 28/28 OK (hs64, 256tok, kda). Gate régression. K>1 non-sup=decode, hors prefill. baseline pp512=93. build: ninja .../libggml-htp-v79.so. P1 en cours.