# 2×RTX 3090 + EPYC box running qwen3.8-flash-next at ~38 tok/s — retained record

> Retained reddit record from reddit:LocalLLaMA.

Canonical: https://highsignal.app/data/records/8e52459f5d98bb65

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:LocalLLaMA)  published 2026-09-12 16:52 ingested 2026-09-13 02:40

[original source ↗](https://www.reddit.com/r/LocalLLaMA/comments/1wehnsd/2rtx_3090_epyc_box_running_qwen38flashnext_at_38/)

## retained text



```
What I have:

\- CPU: EPYC 7551 (32c/64T, Zen 1)

\- Board: Supermicro H11SSL-i (SP3), Rev 2.0

\- RAM: 128 GB DDR4-2133 (all 8 channels full)

\- GPU: 2x RTX 3090 (48 GB total, PCIe 3.0)

\- 1500 W PSU



What I run: 

\- Qwen3-Flash-Next (177B total / \~6B active MoE, IQ4\_XS) on Ilama.cpp. Experts live in system RAM, hot ones cached in VRAM. Single stream = 38 tok/s. Two parallel requests drop to \~4 tok/s each.



Budget: 

\~$800. Realistically that's either one more RTX 3090 or a CPU upgrade (a Zen 2 "Rome" EPYC drops into the same board). A new motherboard is out of budget i think for now. 



Which gives more inference speed for this setup - adding the 3rd 3090, or swapping to a faster/newer CPU? 

And would more/faster RAM matter here? Curious what people running similar rigs have actually measured.

I am also interested in having multiple agents running at the same time, which currently slows it down heavily, so keeping the performance at multiple agents parallel would be a huge boost as well!
```



document fetched 2026-09-13 00:37 · canonical https://reddit.com/r/LocalLLaMA/comments/1wehnsd/2rtx_3090_epyc_box_running_qwen38flashnext_at_38
