Page values for "DeepSeek V4.1 Flash on 8x 8 GB (Zanooda PP8)"
From CMP Wiki
More actions
"Recipe" values
1 row is stored for this page| Field | Field type | Allowed values | Value |
|---|---|---|---|
| page | Page | DeepSeek V4.1 Flash on 8x 8 GB (Zanooda PP8) | |
| model | String | DeepSeek-V4.1-Flash-1017 | |
| quant | String | Official checkpoint as downloaded: MXFP4 routed experts as Marlin W4A16, MXFP8 dense as Marlin W8A16; KV fp8 | |
| engine | String | vLLM PR #56201 head (79a7108d) plus a 20-patch sm_80 series; image zanooda/vllm-sm80-ds41f:v41-sm80 | |
| engine_version | String | recipe 31318fd; patches 0001-0020 | |
| cards | String | 8 GB · 10 GB · both | 8 GB |
| card_count | Integer | 8 | |
| topology | String | PP8, partition 5,5,5,5,5,5,5,5 | |
| context | String | 1,048,576 | |
| decode_tps | Float | 117.0 | |
| prefill_tps | Float | 6,066.0 | |
| vram_per_card | String | 64 GB; weights 35.9 GB per rank (ranks 1-6), 37.6 GB rank 0, 46.3 GB rank 7 | |
| power_cap | String | 180 W (one card unstable at 250 W) | |
| driver | String | ||
| unlock_build | String | cmpunlocker with the BAR1 P2P patches (64 GiB BAR1 per card) | |
| flags | Text | --pipeline-parallel-size 8 --max-model-len 1048576 --gpu-memory-utilization 0.95 --block-size 128 --kv-cache-dtype fp8 --max-num-batched-tokens 4096 --max-num-seqs 32; DSpark 5; VLLM_DSV41_SPREAD_DECODE=2; --engram-config storage cpu | |
| status | String | Proven · Reported · Theory · Contested · Warning | Reported |
| tested_by | String | Zanooda | |
| tested_on | Date | 2026-10-19 | |
| source | String | https://github.com/Zanooda/deepseek-v4.1-flash-cmp170hx | |
| files | Wikitext |