GLM 5.3 on 8x 8 GB (vllm-170hx-glm5)/Confirmation 1
From CMP Wiki
More actions
| Confirms | GLM 5.3 on 8x 8 GB (vllm-170hx-glm5) |
|---|---|
| Result | Reproduced |
| By | @1a2b3c0967 |
| Date | 2026-10-08 |
| Cards | 8 GB |
| Numbers | 8 cards, concurrency 1, KV pool 850k tokens. No tok/s posted; "big performance bump" over his own 8-card build (about 25 tok/s, see GLM 5.3 on 8x 8 GB (Int4-Int8Mix PP8)). |
The build first targeted PP10 and would not run at PP8 for him; it ran once reala added PP8, PP9 and PP10 support the same day. Source: @1a2b3c0967, 2026-10-08
Source: @1a2b3c0967, 2026-10-08