Speed test matrix
Qwen3.8-27B-FP8 on CMP 170HX 64GB
vllm · 0.32.0rc1.dev249+g18305948a · FP86 contexts (0–248K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 128K
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
- 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 Qwen/Qwen3.8-27B-FP8 --max-model-len 262144 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code --language-model-only
Prefill
prompt tok/s by context length8K context1,7174,769 ms TTFT
16K context1,6269,940 ms TTFT
32K context1,53320,957 ms TTFT
64K context1,37346,679 ms TTFT
128K context1,131113,099 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 | 41.3188 ms TTFT | 80.8213 ms TTFT | 156.5417 ms TTFT | 296.9732 ms TTFT | 506.61,297 ms TTFT | 821.92,073 ms TTFT | Over capacity | Over capacity |
| 16K | 39.5451 ms TTFT | 74.8656 ms TTFT | 137.41,057 ms TTFT | 238.61,809 ms TTFT | 362.73,453 ms TTFT | Over capacity | Over capacity | Over capacity |
| 32K | 38.0801 ms TTFT | 69.81,111 ms TTFT | 122.71,754 ms TTFT | 199.63,103 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity |
| 64K | 35.4936 ms TTFT | 61.61,225 ms TTFT | 101.21,820 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity |
| 128K | 31.11,083 ms TTFT | 49.91,322 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity |
| 248K | 23.6Requests queued2,900 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 | 1,717.0 | — | 4,769 ms | — | — | 1 | — | — |
| Prefill | 16K | 1 | Valid | 1,626.0 | — | 9,940 ms | — | — | 1 | — | — |
| Prefill | 32K | 1 | Valid | 1,533.0 | — | 20,957 ms | — | — | 1 | — | — |
| Prefill | 64K | 1 | Valid | 1,373.0 | — | 46,679 ms | — | — | 1 | — | — |
| Prefill | 128K | 1 | Valid | 1,131.0 | — | 113,099 ms | — | — | 1 | — | — |
| Sustained decode | 0 | 1 | Valid | 41.3 | 41.4 | 188 ms | 24.1 ms | 1.0 | 0 | 1,240 | 30.0 s |
| Sustained decode | 0 | 2 | Valid | 80.8 | 40.4 | 213 ms | 24.7 ms | 2.0 | 0 | 2,422 | 30.0 s |
| Sustained decode | 0 | 4 | Valid | 156.5 | 39.2 | 417 ms | 25.5 ms | 4.0 | 0 | 4,692 | 30.0 s |
| Sustained decode | 0 | 8 | Valid | 296.9 | 37.2 | 732 ms | 26.9 ms | 8.0 | 0 | 8,904 | 30.0 s |
| Sustained decode | 0 | 16 | Valid | 506.6 | 31.8 | 1,297 ms | 31.5 ms | 16.0 | 0 | 15,187 | 30.0 s |
| Sustained decode | 0 | 32 | Valid | 821.9 | 25.4 | 2,073 ms | 39.3 ms | 32.0 | 0 | 24,640 | 30.0 s |
| Sustained decode | 0 | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 0 | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 1 | Valid | 39.5 | 39.6 | 451 ms | 25.3 ms | 1.0 | 0 | 1,185 | 30.0 s |
| Sustained decode | 16K | 2 | Valid | 74.8 | 37.2 | 656 ms | 26.8 ms | 2.0 | 0 | 2,242 | 30.0 s |
| Sustained decode | 16K | 4 | Valid | 137.4 | 33.9 | 1,057 ms | 29.5 ms | 4.0 | 0 | 4,120 | 30.0 s |
| Sustained decode | 16K | 8 | Valid | 238.6 | 28.9 | 1,809 ms | 34.6 ms | 8.0 | 0 | 7,152 | 30.0 s |
| Sustained decode | 16K | 16 | Valid | 362.7 | 21.1 | 3,453 ms | 47.4 ms | 16.0 | 0 | 10,880 | 30.0 s |
| Sustained decode | 16K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 1 | Valid | 38.0 | 38.1 | 801 ms | 26.2 ms | 1.0 | 0 | 1,140 | 30.0 s |
| Sustained decode | 32K | 2 | Valid | 69.8 | 34.6 | 1,111 ms | 28.9 ms | 2.0 | 0 | 2,092 | 30.0 s |
| Sustained decode | 32K | 4 | Valid | 122.7 | 30.0 | 1,754 ms | 33.4 ms | 4.0 | 0 | 3,678 | 30.0 s |
| Sustained decode | 32K | 8 | Valid | 199.6 | 23.5 | 3,103 ms | 42.5 ms | 8.0 | 0 | 5,984 | 30.0 s |
| 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 | 35.4 | 35.5 | 936 ms | 28.2 ms | 1.0 | 0 | 1,061 | 30.0 s |
| Sustained decode | 64K | 2 | Valid | 61.6 | 30.7 | 1,225 ms | 32.6 ms | 2.0 | 0 | 1,846 | 30.0 s |
| Sustained decode | 64K | 4 | Valid | 101.2 | 24.8 | 1,820 ms | 40.3 ms | 4.0 | 0 | 3,036 | 30.0 s |
| 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 | 128K | 1 | Valid | 31.1 | 31.2 | 1,083 ms | 32.0 ms | 1.0 | 0 | 933 | 30.0 s |
| Sustained decode | 128K | 2 | Valid | 49.9 | 24.9 | 1,322 ms | 40.1 ms | 2.0 | 0 | 1,496 | 30.0 s |
| Sustained decode | 128K | 4 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 128K | 8 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 128K | 16 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 128K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 128K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 128K | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 248K | 1 | Requests queued | 23.6 | 25.6 | 2,900 ms | 39.1 ms | 0.9 | 0 | 707 | 30.0 s |
| Sustained decode | 248K | 2 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 248K | 4 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 248K | 8 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 248K | 16 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 248K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 248K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 248K | 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.