Drilldown: Recipe - CMP Wiki

Jump to content

Choose a table:

Recipe > page

: GLM 5.3 on 8x 8 GB (vllm-170hx-glm5)

or Qwen3.8 27B on 1x 8 GB (W4A16 vLLM)

& cards: 8 GB

Use the filters below to narrow your results.

GLM-5.3 (full 78-layer DSA MoE, not Flash) (1) · Qwen3.8-27B (dense) (1)

GLM-5.3-Int4-Int8Mix-AWQ-g64; drafter incoai/GLM-5.3-DFlash2 (1) · W4A16 (INT4 weights, Marlin kernels) (1)

Morrowmake/vllm-cmp170hx branch ampere-glm53 plus fail-closed source patches (1) · vLLM (1)

None (1) · vLLM 0.30.1rc1, fork commit 378c37b0 (1)

PP8 (also PP9 and PP10), TP1 (1) · single card (1)

1,048,576 per request (1) · 262,144 native (YaRN to about 500k in the recipe's long variant) (1)

48.7 (1) · 170 (1)

None (1) · 1,833 (1)

64 GB (1) · 64 GiB, GPU_UTIL 0.96 at PP8 (1)

None (1) · 170 W (author's 10-card run) (1)

None (2)

None (2)

Proven (1) · Reported (1)

kellertreppe123 (1) · reala7096 (1)

September 29, 2026 (1) · October 5, 2026 (1)

https://github.com/arvkmr/vllm-170hx-glm5 (1) · https://www.reddit.com/r/LocalLLM/comments/1w2f03e/qwen3827b_w4a16_on_one_64_gb_cmp_170hx_16070/ (1)

Showing below up to 2 results in range #1 to #2.

G

GLM 5.3 on 8x 8 GB (vllm-170hx-glm5)

Q

Qwen3.8 27B on 1x 8 GB (W4A16 vLLM)

Retrieved from "https://cmp.wiki/wiki/Special:Drilldown/Recipe"