Speed test matrix

DeepSeek-R1-Distill-Qwen-32B-FP8-dynamic on CMP 170HX 64GB

vllm · 0.32.0rc1.dev249+g18305948a · FP8 dynamic5 contexts (0–88K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 88K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 30, 2026, 10:33 UTC

Hardware

Type: Discrete GPU
GPU: CMP 170HX
VRAM: 64 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: FP8 dynamic
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
# 1x 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 RedHatAI/DeepSeek-R1-Distill-Qwen-32B-FP8-dynamic --max-model-len 98304 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code

Prefill

prompt tok/s by context length

8K context1,8564,415 ms TTFT

16K context1,6669,742 ms TTFT

32K context1,43622,513 ms TTFT

64K context1,12057,624 ms TTFT

88K context960.092,361 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	32.687 ms TTFT	63.9104 ms TTFT	126.0112 ms TTFT	244.1149 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
16K	29.0188 ms TTFT	53.1250 ms TTFT	91.8350 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	26.3323 ms TTFT	45.0395 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
64K	22.1519 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
88K	19.7769 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity

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	1,856.0	—	4,415 ms	—	—	1	—	—
Prefill	16K	1	Valid	1,666.0	—	9,742 ms	—	—	1	—	—
Prefill	32K	1	Valid	1,436.0	—	22,513 ms	—	—	1	—	—
Prefill	64K	1	Valid	1,120.0	—	57,624 ms	—	—	1	—	—
Prefill	88K	1	Valid	960.0	—	92,361 ms	—	—	1	—	—
Sustained decode	0	1	Valid	32.6	32.6	87 ms	30.7 ms	1.0	0	977	30.0 s
Sustained decode	0	2	Valid	63.9	32.0	104 ms	31.2 ms	2.0	0	1,914	30.0 s
Sustained decode	0	4	Valid	126.0	31.6	112 ms	31.6 ms	4.0	0	3,776	30.0 s
Sustained decode	0	8	Valid	244.1	30.7	149 ms	32.6 ms	8.0	0	7,320	30.0 s
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	29.0	29.1	188 ms	34.4 ms	1.0	0	870	30.0 s
Sustained decode	16K	2	Valid	53.1	26.6	250 ms	37.6 ms	2.0	0	1,592	30.0 s
Sustained decode	16K	4	Valid	91.8	23.0	350 ms	43.5 ms	4.0	0	2,752	30.0 s
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	Valid	26.3	26.3	323 ms	38.0 ms	1.0	0	788	30.0 s
Sustained decode	32K	2	Valid	45.0	22.5	395 ms	44.3 ms	2.0	0	1,350	30.0 s
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	—	—	—	—	—	—	—	—
Sustained decode	64K	1	Valid	22.1	22.1	519 ms	45.2 ms	1.0	0	662	30.0 s
Sustained decode	64K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	88K	1	Valid	19.7	19.8	769 ms	50.6 ms	1.0	0	591	30.0 s
Sustained decode	88K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	88K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	88K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	88K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	88K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	88K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	88K	128	Over capacity	—	—	—	—	—	—	—	—

Over capacity: not run because this many requests at this context would exceed the server's KV cache.