Page values for "Qwen3.8 Flash Next on 2x 8 GB (Morrowmake EXL3)"
From CMP Wiki
More actions
"Recipe" values
1 row is stored for this page| Field | Field type | Allowed values | Value |
|---|---|---|---|
| page | Page | Qwen3.8 Flash Next on 2x 8 GB (Morrowmake EXL3) | |
| model | String | Qwen3.8-Flash-Next | |
| quant | String | EXL3 5.05 bpw; PLE (n-gram) table read from SSD | |
| engine | String | Morrowmake/vllm-cmp170hx (EXL3 path) | |
| engine_version | String | tag qwen38-recipe-1.0.0, engine 18305948ac | |
| cards | String | 8 GB · 10 GB · both | 8 GB |
| card_count | Integer | 2 | |
| topology | String | TP2 with expert parallel | |
| context | String | 262,144 | |
| decode_tps | Float | 152.6 | |
| prefill_tps | Float | 2,651.0 | |
| vram_per_card | String | 64 GiB, --gpu-memory-utilization 0.95 | |
| power_cap | String | ||
| driver | String | ||
| unlock_build | String | ||
| flags | Text | --quantization exl3 --load-format exl3 --tensor-parallel-size 2 --enable-expert-parallel --engram-config cpu_offload false, VLLM_EXL3_PLE_PLACEMENT=ssd, cudagraph PIECEWISE, MTP 3 | |
| status | String | Proven · Reported · Theory · Contested · Warning | Reported |
| tested_by | String | voidgiraffe | |
| tested_on | Date | 2026-10-16 | |
| source | String | https://github.com/Morrowmake/qwen38-flash-next-cmp170hx-recipe | |
| files | Wikitext |