Speed test matrix

DeepSeek-R1-Distill-Qwen-14B-FP8-dynamic on CMP 170HX 64GB

vllm · 0.32.0rc1.dev249+g18305948a · FP8 dynamic5 contexts (0–120K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 120K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 30, 2026, 21:21 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-14B-FP8-dynamic --max-model-len 131072 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code

Prefill

prompt tok/s by context length

8K context3,9102,096 ms TTFT

16K context3,3864,794 ms TTFT

32K context2,71811,893 ms TTFT

64K context1,94733,141 ms TTFT

120K context1,29493,367 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	65.363 ms TTFT	126.4123 ms TTFT	244.8122 ms TTFT	461.4155 ms TTFT	844.7191 ms TTFT	Over capacity	Over capacity	Over capacity
16K	55.3173 ms TTFT	97.9224 ms TTFT	162.4323 ms TTFT	238.5478 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
32K	48.1280 ms TTFT	79.1382 ms TTFT	117.2551 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
64K	38.4487 ms TTFT	55.6676 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
120K	28.0959 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	3,910.0	—	2,096 ms	—	—	1	—	—
Prefill	16K	1	Valid	3,386.0	—	4,794 ms	—	—	1	—	—
Prefill	32K	1	Valid	2,718.0	—	11,893 ms	—	—	1	—	—
Prefill	64K	1	Valid	1,947.0	—	33,141 ms	—	—	1	—	—
Prefill	120K	1	Valid	1,294.0	—	93,367 ms	—	—	1	—	—
Sustained decode	0	1	Valid	65.3	65.5	63 ms	15.3 ms	1.0	0	1,958	30.0 s
Sustained decode	0	2	Valid	126.4	63.5	123 ms	15.7 ms	2.0	0	3,790	30.0 s
Sustained decode	0	4	Valid	244.8	61.8	122 ms	16.2 ms	4.0	0	7,344	30.0 s
Sustained decode	0	8	Valid	461.4	58.5	155 ms	17.1 ms	8.0	0	13,838	30.0 s
Sustained decode	0	16	Valid	844.7	54.2	191 ms	18.5 ms	16.0	0	25,328	30.0 s
Sustained decode	0	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	55.3	55.5	173 ms	18.0 ms	1.0	0	1,660	30.0 s
Sustained decode	16K	2	Valid	97.9	49.1	224 ms	20.3 ms	2.0	0	2,938	30.0 s
Sustained decode	16K	4	Valid	162.4	40.7	323 ms	24.6 ms	4.0	0	4,868	30.0 s
Sustained decode	16K	8	Valid	238.5	29.8	478 ms	33.6 ms	8.0	0	7,152	30.0 s
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	48.1	48.3	280 ms	20.7 ms	1.0	0	1,443	30.0 s
Sustained decode	32K	2	Valid	79.1	39.6	382 ms	25.2 ms	2.0	0	2,372	30.0 s
Sustained decode	32K	4	Valid	117.2	29.3	551 ms	34.1 ms	4.0	0	3,516	30.0 s
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	38.4	38.5	487 ms	26.0 ms	1.0	0	1,151	30.0 s
Sustained decode	64K	2	Valid	55.6	27.8	676 ms	35.9 ms	2.0	0	1,668	30.0 s
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	120K	1	Valid	28.0	28.2	959 ms	35.5 ms	1.0	0	840	30.0 s
Sustained decode	120K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	128	Over capacity	—	—	—	—	—	—	—	—

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