Page values for "GLM 5.3 Flash on 4x 8 GB"
From CMP Wiki
More actions
"Recipe" values
1 row is stored for this page| Field | Field type | Allowed values | Value |
|---|---|---|---|
| page | Page | GLM 5.3 Flash on 4x 8 GB | |
| model | String | GLM-5.3-Flash (320B MoE, 18B active) | |
| quant | String | W4A16 (INT4 weights, group size 128, FP16 activations): canada-quant/GLM-5.3-Flash-W4A16-MTP | |
| engine | String | Morrowmake/vllm-cmp170hx (vLLM fork, container image) | |
| engine_version | String | recipe 1.8.0, engine 819146023f | |
| cards | String | 8 GB · 10 GB · both | 8 GB |
| card_count | Integer | 4 | |
| topology | String | TP4 (default) or PP4 | |
| context | String | 262,144 | |
| decode_tps | Float | 442.1 | |
| prefill_tps | Float | 3,161.0 | |
| vram_per_card | String | 64 GiB exposed, --gpu-memory-utilization 0.95 | |
| power_cap | String | 180 W | |
| driver | String | NVIDIA 580 or newer; P2P needs the Morrowmake/cmpunlocker build | |
| unlock_build | String | cmpunlocker (64 GiB, 74 SMs); Morrowmake/cmpunlocker for P2P | |
| flags | Text | LAYOUT=tp4 or LAYOUT=pp4, P2P=auto, --max-num-seqs 8, --gpu-memory-utilization 0.95, --max-num-batched-tokens 3463 (TP4) or 2312 (PP4) | |
| status | String | Proven · Reported · Theory · Contested · Warning | Proven |
| tested_by | String | voidgiraffe | |
| tested_on | Date | 2026-10-16 | |
| source | String | https://github.com/Morrowmake/glm53-flash-cmp170hx-recipe | |
| files | Wikitext |