GLM 5.2 Hits 2626 tok/s on a Single AMD MI355X Node
burny_tech · x · 2026-07-05
According to waferai (which hit #2 on Hacker News), engineers successfully deployed and served GLM 5.2 on AMD MI355X. The single-node throughput reached 2626 tok/s, with about 213 tok/s per user, demonstrating the potential of non-Nvidia hardware for open-source large model inference.
Related event: GLM 5.2 Achieves 2626 tok/s Inference on AMD MI355X(2 posts)→
More from Infra
- Engram's random reads don't suit SSDs; CPU-memory over NVLink could serve all 72 GPUs — bookwormengr · 2026-09-11
- 80% of the DIY LLM inference hype posters have already quit — it's brutally hard systems work — abhijithneil · 2026-09-11
- Hugging Face's Ultra Scale Playbook: a free book on training LLMs on GPU clusters — mdancho84 · 2026-09-11
- Is inference latency becoming the biggest bottleneck for production AI agents? — Euphoric_Sea632 · 2026-09-11
- LLM Serving Metrics Thread: Why TPOT and Uptime Make or Break User Experience — abhijithneil · 2026-09-11
- PlanetScale launches sharded Postgres: 768 servers acting as one, 1PB scale — dhruv2038 · 2026-09-11