Speed test matrix
Muse-Glimmer-30B-INT4 on RTX PRO 6000 Blackwell Server Edition 96GB
vllm · 0.30.1rc1.dev452+g73a583112 · INT45 contexts (8K–123K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 123K, 128K
Hardware
- Type
- Discrete GPU
- GPU
- RTX PRO 6000 Blackwell Server Edition
- VRAM
- 96 GB
- OS
- linux
Engine
- Engine
- vllm
- Version
- 0.30.1rc1.dev452+g73a583112
- Quantization
- INT4
- Backend
- cuda
- Repository
- github.com/vllm-project/vllm
- Build
- cuda-13.2
- Commit
73a5831127a9
Run
- Served model
- muse-glimmer
- Revision
- main
- Benchmark tool
- llm-inference-bench v0.7.6
- Decode test
- Fixed duration
- Prefill mode
- Measured during decode warm-up
- Time per measurement
- 30 s
- Max output
- 4,096 tokens
Launch command
env MAX_JOBS=2 FLASHINFER_NVCC_THREADS=1 TORCHINDUCTOR_COMPILE_THREADS=1 "$HOME/env_rtxpro6k-redhatai-vllm-73a5831127a9d2b8/bin/vllm" serve RedHatAI/Muse-Glimmer-30B-W4A16 --served-model-name muse-glimmer --tensor-parallel-size 1 --reasoning-parser muse_glimmer --enable-auto-tool-choice --tool-call-parser muse_glimmer --max-model-len 131072 --gpu-memory-utilization 0.83 --host 0.0.0.0 --port 8000 --api-key <redacted> --generation-config auto
Prefill
prompt tok/s by context length8K context5,3881,521 ms TTFT
16K context5,3213,041 ms TTFT
32K context5,7735,573 ms TTFT
64K context5,50211,660 ms TTFT
123K context4,85925,338 ms TTFT
128K context4,88726,212 ms TTFT
Sustained decode
total tok/s · rows are context length, columns are simultaneous requests| Context | 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128 |
|---|---|---|---|---|---|---|---|---|
| 8K | Looped | Looped | Looped | Looped | Looped | 1,082.2373 ms TTFT | 1,630.7579 ms TTFT | 1,794.91,144 ms TTFT |
| 16K | Looped | Looped | Looped | Looped | Looped | 1,106.0496 ms TTFT | 1,553.9903 ms TTFT | 1,722.01,762 ms TTFT |
| 32K | 63.1 | Looped | Looped | Looped | Looped | 999.7917 ms TTFT | 1,434.91,565 ms TTFT | Over capacity |
| 64K | 61.6182 ms TTFT | Looped | Looped | Looped | Looped | 892.31,298 ms TTFT | Over capacity | Over capacity |
| 123K | 59.8 | Looped | 172.2 | Looped | 496.41,696 ms TTFT | Over capacity | Over capacity | Over capacity |
Looped: the model kept repeating the same text, so the benchmark discarded this measurement. 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 | 5,388.0 | — | 1,521 ms | — | — | 1 | — | — |
| Prefill | 16K | 1 | Valid | 5,321.0 | — | 3,041 ms | — | — | 1 | — | — |
| Prefill | 32K | 1 | Valid | 5,773.0 | — | 5,573 ms | — | — | 1 | — | — |
| Prefill | 64K | 1 | Valid | 5,502.0 | — | 11,660 ms | — | — | 1 | — | — |
| Prefill | 123K | 1 | Valid | 4,859.0 | — | 25,338 ms | — | — | 1 | — | — |
| Prefill | 128K | 1 | Valid | 4,887.0 | — | 26,212 ms | — | — | 1 | — | — |
| Sustained decode | 8K | 1 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 2 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 4 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 8 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 16 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 32 | Valid | 1,082.2 | 37.8 | 373 ms | 26.5 ms | 32.0 | 0 | 32,467 | 30.0 s |
| Sustained decode | 8K | 64 | Valid | 1,630.7 | 25.6 | 579 ms | 39.1 ms | 64.0 | 0 | 48,922 | 30.0 s |
| Sustained decode | 8K | 128 | Valid | 1,794.9 | 14.2 | 1,144 ms | 70.6 ms | 128.0 | 0 | 53,848 | 30.0 s |
| Sustained decode | 16K | 1 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 16K | 2 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 16K | 4 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 16K | 8 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 16K | 16 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 16K | 32 | Valid | 1,106.0 | 36.5 | 496 ms | 27.4 ms | 32.0 | 0 | 33,181 | 30.0 s |
| Sustained decode | 16K | 64 | Valid | 1,553.9 | 24.3 | 903 ms | 41.2 ms | 64.0 | 0 | 46,616 | 30.0 s |
| Sustained decode | 16K | 128 | Valid | 1,722.0 | 13.4 | 1,762 ms | 74.4 ms | 128.0 | 0 | 51,706 | 30.0 s |
| Sustained decode | 32K | 1 | Valid | 63.1 | — | — | — | 1.0 | 0 | 686 | 10.9 s |
| Sustained decode | 32K | 2 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 32K | 4 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 32K | 8 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 32K | 16 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 32K | 32 | Valid | 999.7 | 34.1 | 917 ms | 29.3 ms | 32.0 | 0 | 29,991 | 30.0 s |
| Sustained decode | 32K | 64 | Valid | 1,434.9 | 22.3 | 1,565 ms | 44.8 ms | 64.0 | 0 | 43,047 | 30.0 s |
| Sustained decode | 32K | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 64K | 1 | Valid | 61.6 | 61.7 | 182 ms | 16.2 ms | 1.0 | 0 | 1,848 | 30.0 s |
| Sustained decode | 64K | 2 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 64K | 4 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 64K | 8 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 64K | 16 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 64K | 32 | Valid | 892.3 | 29.9 | 1,298 ms | 33.4 ms | 32.0 | 0 | 26,768 | 30.0 s |
| Sustained decode | 64K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 64K | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 123K | 1 | Valid | 59.8 | — | — | — | 1.0 | 0 | 783 | 13.1 s |
| Sustained decode | 123K | 2 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 123K | 4 | Valid | 172.2 | — | — | — | 4.0 | 0 | 2,758 | 16.0 s |
| Sustained decode | 123K | 8 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 123K | 16 | Valid | 496.4 | 37.0 | 1,696 ms | 27.1 ms | 16.0 | 0 | 14,887 | 30.0 s |
| Sustained decode | 123K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 123K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 123K | 128 | Over capacity | — | — | — | — | — | — | — | — |
Looped: the model kept repeating the same text, so the benchmark discarded this measurement. Over capacity: not run because this many requests at this context would exceed the server's KV cache.