CachyOS turns out slower than Windows for ROCm video generation on RX 9070XT
Sofa-Sleuth · reddit · 2026-10-11
A user spent a day setting up a CachyOS + Windows dual boot on a new NVMe for an RX 9070XT, installed ROCm 10.1, Python 3.13 and PyTorch 2.14, and found CachyOS consistently slower than Windows for MiniMax H3 video generation.
- Same workload: 0.6MP, 20–25 step, 12-second image-to-video
- Windows: 40–50 seconds per step
- CachyOS: 45–80 seconds per step
The result challenges the common assumption that Linux + ROCm outperforms Windows for local AMD inference. The author is asking the community for comparable experiences.
More from Infra
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Jensen Huang lays out NVIDIA's full Vera Rubin agentic system: NVL72, MGX, CMX and RTX — rohanpaul_ai · 2026-10-23
- ONNX to Transformers.js: running models directly in your browser — JosephJacks_ · 2026-10-23
- MiniMax M3.1 Flash free for 3 days on 0G, with 1M context and 512K output — MiniMax_AI · 2026-10-23
- "We bet on Cloudflare and it paid off": R2 and D1 at billion scale — dinasaur_404 · 2026-10-23
- Public data is drying up; private training data is turning into a browsable catalog — SucceededMind · 2026-10-23