Drilldown: Recipe

DeepSeek V4.1 Flash on 4x 8 GB (GNmac) (1) · DeepSeek V4.1 Flash on 4x 8 GB plus RTX PRO 6000 (ExLlamaV3 TP5) (1)

DeepSeek-V4.1-Flash-1017 (2)

EXL3 3.5 bpw (LibertAIDAI; routed experts 4 bpw on layers 0-9 and 30-39, 3 bpw on 10-29; attention and shared experts 6 bpw) plus the official Engram shards (1) · Hybrid: routed experts EXL3 (mostly 3 bpw, 2 bpw on layers 18-22) from LibertAI's 2.9 bpw pack; attention, shared experts, indexer and DSpark drafter official FP8; KV fp8_ds_mla (1)

ExLlamaV3 fork (anoane/exllamav3-anemone) + TabbyAPI fork (anoane/tabbyAPI) (1) · vLLM with a pinned sm_80 patch stack (built locally from source) (1)

recipe 1.2.0 (1) · tag recipe-09 (309ef6d1); TabbyAPI deepseek-v41 (ce250ff) (1)

8 GB · Other · None

4 (2)

TP4 (default) or PP4 for x1/x4 links (1) · TP5 across four CMP 170HX and one RTX PRO 6000 Blackwell (native backend, reduces through host memory); MTP draft on the PRO (1)

1,048,576 FP16 (1) · 262,144 (up to 1,048,576) (1)

99.9 (1) · 258 (1)

2,116 (1) · 2,416 (1)

64 GiB, GPU_MEMORY_UTILIZATION 0.95 (1) · gpu_split 58 GB per CMP, 70 GB on the PRO (1)

None · 180 W · 200 W · 140 W · 140 W (200 W observed-draw abort guard) · 150 W per CMP (PRO 400 W) · 170 W (author's 10-card run) · 180 W (author's cards); about 170 W with a modest memory OC in his Discord numbers · 180 W (nvidia-smi -pl 180) · 180 W (one card unstable at 250 W) · Other

CUDA 13 capable; measured on 610.43.02 (1) · NVIDIA open 610.57.04 patched with Morrowmake/cmpunlocker 1f2c498 (profile 8gb) (1)

cmpunlocker (64 GiB, 74 SMs) with a P2P-capable driver for TP4 (1) · Morrowmake/cmpunlocker 1f2c498; modprobe RmForceEnableGen2=1, RMPcieLinkSpeed=0x1, ForceP2P=0x11 (1)

Proven (1) · Reported (1)

anoane (1) · greencandle9753 (1)

October 20, 2026 (1) · October 26, 2026 (1)

https://github.com/anoane/09-deepseek-v41-flash-1017-exl3-tp5-1m (1) · https://github.com/GNmac-lab/deepseek-v41-flash-1017-cmp170hx (1)