Hands-on: AMD ROCm 10.2 delivers 2.2x token throughput on Qwen3.8-27B
AnushElangovan · x · 2026-11-21
A tester running a local data center reports AMD's newly released ROCm 10.2 stack — positioned as a CUDA alternative — is "legit": on AMD GPUs with multi-lane concurrency, Qwen3.8-27B saw 2.2x token throughput. Tests on Qwen3.8-Flash-Next are underway with early numbers looking "ridiculously good." If reproducible, it's a notable step for AMD on LLM inference.
More from Infra
- OpenVINO: the open-source toolkit for running AI on Intel CPUs, GPUs and NPUs — abhishek__AI · 2026-11-21
- AWS Lambda now supports GPUs via EC2 capacity providers — samgoodwin89 · 2026-11-21
- 27,262 thinking tokens, zero answer: Google auto-billed 23 empty Vertex AI crashes via HTTP 200 — Plenty-Jackfruit-218 · 2026-11-21
- Hy4 Flash reads 1M tokens in 44 min on Mac Studio M3 Ultra, 81 min on DGX Spark — AIFlow_ML · 2026-11-21
- local-ai-registry: one battle-tested local LLM recipe per GPU, with evidence attached — LiTianleli · 2026-11-21
- Sous Labs open-sources Songe, a local fine-tuning engine that trains on your own conversations — alexcovo_eth · 2026-11-21