Speed test matrix

GLM-4.7-218B-W4A16 on CMP 170HX 64GB ×2

vllm · 0.32.0rc1.dev249+g18305948a · AutoRound W4A163 contexts (0–32K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 31, 2026, 14:54 UTC

Hardware

Type: Discrete GPU
GPU: CMP 170HX
GPU count: 2
VRAM: 64 GB (2 × 32 GB)
CPU: 2x Intel Xeon E5-2630 v3
RAM: 64 GB
OS: Ubuntu 26.04.1

Engine

Engine: vllm
Version: 0.32.0rc1.dev249+g18305948a
Quantization: AutoRound W4A16
Backend: cuda
Repository: github.com/vllm-project/vllm
Build: ghcr.io/morrowmake/vllm-cmp170hx@sha256:5b45c4d6feafefe2eba962375b0a9a0b64fd22b9e05abb4d1966f14a32a2db79

Run

Served model: bench
Revision: main
Benchmark tool: llm-inference-bench v0.8.0
Decode test: Fixed duration
Prefill mode: Measured during decode warm-up
Time per measurement: 30 s
Max output: 8,192 tokens

Launch command

# Untuned out-of-the-box baseline. Spec decode: none. KV cache: int4_per_token_head. Layout: PP2 (stock vLLM flag)
# 2x NVIDIA CMP 170HX 64GB (cmpunlocker), 180 W limit, PCIe Gen2 x4 | 2x Intel Xeon E5-2630 v3 | 64 GB DDR4 | Ubuntu 26.04.1, NVIDIA driver 610.57.04 | vLLM: Morrowmake vllm-cmp170hx image
vllm serve 0xSero/GLM-4.7-218B-W4A16 --max-model-len 40960 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code --pipeline-parallel-size 2 --kv-cache-dtype int4_per_token_head --attention-backend TRITON_ATTN

Prefill

prompt tok/s by context length

8K context593.013,790 ms TTFT

16K context411.039,396 ms TTFT

32K context248.0130,176 ms TTFT

Sustained decode

total tok/s · rows are context length, columns are simultaneous requests

Context	1	2	4	8	16	32	64	128
0	33.3204 ms TTFT	50.4336 ms TTFT	88.5325 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
16K	20.3404 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	14.7Requests queued661 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity

Requests queued: the server could not run all requests at once, so some waited in a queue. Over capacity: not run because this many requests at this context would exceed the server's KV cache.

TTFT is measured after the benchmark warms the server's prompt cache, so it mostly excludes prefill.

All measurements

Latencies are medians. Total tok/s is output across all simultaneous requests.

Phase	Context	Requests at once	Status	Total tok/s	Per request tok/s	TTFT	Time between tokens	Average active requests	Completed requests	Output tokens	Measured for
Prefill	8K	1	Valid	593.0	—	13,790 ms	—	—	1	—	—
Prefill	16K	1	Valid	411.0	—	39,396 ms	—	—	1	—	—
Prefill	32K	1	Valid	248.0	—	130,176 ms	—	—	1	—	—
Sustained decode	0	1	Valid	33.3	33.4	204 ms	29.9 ms	1.0	0	998	30.0 s
Sustained decode	0	2	Valid	50.4	25.3	336 ms	39.6 ms	2.0	0	1,512	30.0 s
Sustained decode	0	4	Valid	88.5	22.2	325 ms	45.0 ms	4.0	0	2,656	30.0 s
Sustained decode	0	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	20.3	20.4	404 ms	49.1 ms	1.0	0	610	30.0 s
Sustained decode	16K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	1	Requests queued	14.7	14.8	661 ms	67.7 ms	1.0	0	441	29.9 s
Sustained decode	32K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	128	Over capacity	—	—	—	—	—	—	—	—

Requests queued: the server could not run all requests at once, so some waited in a queue. Over capacity: not run because this many requests at this context would exceed the server's KV cache.