Drilldown: Recipe

DeepSeek V4.1 Flash on 8x 8 GB (Zanooda PP8) (1) · GLM 5.3 Flash on 8x 8 GB (promisezackr PP8 NVFP4 DFlash2) (1) · Qwen3.8 Flash Next on 1x 8 GB (iIIusi0n SSD PLE) (1) · Qwen3.8 Flash Next on 2x 8 GB (DUFRelic PP2) (1) · Qwen3.8 Flash Next on 4x 8 GB (1)

DeepSeek-V4.1-Flash-1017 (1) · GLM-5.3-Flash (1) · Qwen3.8-Flash-Next (3)

AutoRound 3 bpw + MTP: klee100/Qwen3.8-Flash-Next-AutoRound-3bpw-MTP; 95.37 GiB BF16 PLE table on SSD (1) · AWQ W4A16: wtdcode/Qwen3.8-Flash-Next-AWQ-W4A16 (1) · LibertAIDAI/GLM-5.3-Flash-NVFP4 (ModelOpt NVFP4, routed experts only, 181 GiB), run through Marlin W4A16 kernels; GLM-5.3-Flash DFlash2 drafter (5 sliding-window layers) (1) · Official checkpoint as downloaded: MXFP4 routed experts as Marlin W4A16, MXFP8 dense as Marlin W8A16; KV fp8 (1) · W4A16 INT4 g128 with the PLE table repacked INT4 g32: zankich/Qwen3.8-Flash-Next-W4A16-Merlin-INT4PLE-FP8KV; KV BF16 (1)

vLLM a5a30471 plus a 13-file downstream patch (native Linux AIO PLE reader) (1) · vLLM PR #56201 head (79a7108d) plus a 20-patch sm_80 series; image zanooda/vllm-sm80-ds41f:v41-sm80 (1) · vLLM v0.30.0 + zankich INT4-PLE plugin + six hash-guarded patch layers (1) · vLLM: official vllm/vllm-openai:glm53-flash image plus promisezackr's 24 patches (SM8x base from 344303947/vllm170hx, pipeline parallel for GLM-5.3-Flash, DFlash2 under PP8) (1) · wtdcode/vllm-backport (image lazymio/vllm-backport:latest-sm80) (1)

0.29.1rc1.dev402+ga5a30471f.ple1 (1) · latest-sm80 tag as of 2026-10-16 (1) · recipe 31318fd; patches 0001-0020 (1) · v0.1.dev20051+g487ecf187 (the image's vLLM); repo commit 90ec72e (1) · vLLM v0.30.0 (ced6857a); plugin zankich/vllm v0.29.0z (6609bfef) (1)

8 GB (5)

1 (1) · 2 (1) · 4 (1) · 8 (2)

PP2, MTP depth 3 (1) · PP8 (VLLM_PP_LAYER_PARTITION=7,5,7,5,7,5,7,2) with DFlash2, 7 draft tokens; PCIe Gen2 x4, no P2P; host 2x Xeon E5-2690 v4, 96 GB DDR4 (1) · PP8, partition 5,5,5,5,5,5,5,5 (1) · single card (1) · TP4 with expert parallel, MTP 3 (1)

262,144 · 1,048,576 · None · 1,000,000 (YaRN x4 from 262,144) · 1,047,736 verified one-shot; 1,002,852 over a 405-turn chat with DSV4_LOGITS_ROW_CHUNK=64 · 1,048,576 (KV pool 1,632,827 tokens at utilisation 0.97; an 896K-token prompt passed) · 1,048,576 (KV pool 6.62M tokens at utilization 0.90, about 6.3 concurrent 1M requests) · 1,048,576 (YaRN x4) · 1,048,576 FP16 · 1,048,576 per request · 262,144 (256K in the post) · 262,144 (up to 1,048,576) · 262,144 (weights address 1,048,576) · 262,144 advertised (tested through 32K) · 262,144 native (YaRN to about 500k in the recipe's long variant) · 262,144 native; 524,288 with YaRN 2.0; 1,048,576 with YaRN 4.0 (needle verified on PP3) · 392,960 (profile q8_384k) · 392,960 per request; 655,360-token Q8 KV pool shared by two requests · 524,288 · 691,968 (largest stable context at --gpu-memory-utilization 0.965) · about 1,000,000 (8-bit KV cache) · Other

110.99 (1) · 116.49 (1) · 117 (1) · 123 (1) · 133.39 (1)

None (1) · 2,101.4 (1) · 2,636 (1) · 6,066 (1) · 6,213.66 (1)

64 GB (1) · 64 GB, --gpu-memory-utilization 0.90 (1) · 64 GB; weights 35.9 GB per rank (ranks 1-6), 37.6 GB rank 0, 46.3 GB rank 7 (1) · 64 GiB, --gpu-memory-utilization 0.90 (1) · 64 GiB, --gpu-memory-utilization 0.96 (1)

None (2) · 140 W (200 W observed-draw abort guard) (1) · 180 W (1) · 180 W (one card unstable at 250 W) (1)

None (3) · 610.43.03 (1) · NVIDIA 610.43.02 (1)

None (1) · amoghmunikote/cmpunlocker 6c442eeb (patch 0016 SM reconfig, 74 SMs) plus Bayley P2P patches 0011/0013/0015 (1) · cmpunlocker (64 GiB) (1) · cmpunlocker VRAM unlock (64 GB per card) (1) · cmpunlocker with the BAR1 P2P patches (64 GiB BAR1 per card) (1)

Proven (1) · Reported (4)

dufrelic (1) · iIIusi0n (1) · matstatman (1) · promisezackr (1) · Zanooda (1)

August 2026 (1) · September 2026 (1) · October 2026 (3)

https://discord.com/channels/1518634379604136178/1528826330454429706/1560602001908170753 (1) · https://github.com/DUFRelic/qwen38-flash-next-recipe (1) · https://github.com/iIIusi0n/qwen3.8-flash-next-cmp170hx (1) · https://github.com/promisezackr/glm53-flash-170hx-pp8 (1) · https://github.com/Zanooda/deepseek-v4.1-flash-cmp170hx (1)