Speed test matrix

granite-4.2-8b on RTX 3090 24GB

vllm · 0.32.0 · BF165 contexts (0–96K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 96K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 31, 2026, 08:40 UTC

Hardware

Type: Discrete GPU
GPU: RTX 3090
VRAM: 24 GB
CPU: Intel Core i7-9700K (8C)
RAM: 32 GB
OS: Ubuntu 24.04.5

Engine

Engine: vllm
Version: 0.32.0
Quantization: BF16
Backend: cuda
Repository: github.com/vllm-project/vllm
Build: vllm/vllm-openai:v0.32.0-x86_64

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
# 1x RTX 3090 24GB (Zagitta/gpu-unlock), 420 W limit, PCIe Gen3 x16 | Intel Core i7-9700K 8C | 32 GB DDR4 | Ubuntu 24.04.5, NVIDIA driver 580.178.04
vllm serve ibm-granite/granite-4.2-8b --max-model-len 106496 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code  --kv-cache-dtype int4_per_token_head --attention-backend TRITON_ATTN

Prefill

prompt tok/s by context length

8K context3,5182,343 ms TTFT

16K context2,4186,748 ms TTFT

32K context1,51721,403 ms TTFT

64K context861.075,242 ms TTFT

96K context599.0161,933 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	44.441 ms TTFT	84.945 ms TTFT	165.058 ms TTFT	316.462 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
16K	35.5110 ms TTFT	65.1157 ms TTFT	105.0243 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	29.6231 ms TTFT	52.6296 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
64K	22.1360 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
96K	17.7Requests queued530 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	3,518.0	—	2,343 ms	—	—	1	—	—
Prefill	16K	1	Valid	2,418.0	—	6,748 ms	—	—	1	—	—
Prefill	32K	1	Valid	1,517.0	—	21,403 ms	—	—	1	—	—
Prefill	64K	1	Valid	861.0	—	75,242 ms	—	—	1	—	—
Prefill	96K	1	Valid	599.0	—	161,933 ms	—	—	1	—	—
Sustained decode	0	1	Valid	44.4	44.5	41 ms	22.5 ms	1.0	0	1,333	30.0 s
Sustained decode	0	2	Valid	84.9	42.5	45 ms	23.5 ms	2.0	0	2,546	30.0 s
Sustained decode	0	4	Valid	165.0	41.5	58 ms	24.1 ms	4.0	0	4,947	30.0 s
Sustained decode	0	8	Valid	316.4	39.9	62 ms	25.1 ms	8.0	0	9,488	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	35.5	35.5	110 ms	28.1 ms	1.0	0	1,065	30.0 s
Sustained decode	16K	2	Valid	65.1	32.5	157 ms	30.7 ms	2.0	0	1,950	30.0 s
Sustained decode	16K	4	Valid	105.0	26.3	243 ms	38.1 ms	4.0	0	3,147	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	29.6	29.6	231 ms	33.8 ms	1.0	0	887	30.0 s
Sustained decode	32K	2	Valid	52.6	26.3	296 ms	38.0 ms	2.0	0	1,578	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	360 ms	45.2 ms	1.0	0	663	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	96K	1	Requests queued	17.7	17.7	530 ms	56.5 ms	1.0	0	530	30.0 s
Sustained decode	96K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	96K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	96K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	96K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	96K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	96K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	96K	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.