AMD ROCm 10.1 fixes major issues: LLaMA.cpp runs flawlessly on RDNA2
smellof · reddit · 2026-08-21
A user reports a breakthrough experience using the TheRock native "multi-arch" release to install ROCm 10.1 on Fedora 44. Previously criticized for compatibility, ROCm successfully compiled and ran llama.cpp on RDNA2 graphics cards without needing overrides or manual source compilation of HIP packages.
More from Infra
- Haight Street Compute Market: a one-day trading floor for GPU deals in SF — ctjlewis · 2026-10-16
- Temporal Serverless Workers on Cloud Run: queue-driven autoscaling, scale to zero — rseroter · 2026-10-16
- LMA brings day-0 model support to TPU for the first time, spanning NVIDIA, AMD, Intel via SGLang — ying11231 · 2026-10-16
- Semtech's analyst-day slide cleanly explains the four signal-integrity impairments — BenBajarin · 2026-10-16
- Fortytwo Engine boosts Qwen3.8-27B accuracy 7% and cuts tokens 25% via Swarm Inference — rohanpaul_ai · 2026-10-16
- Berkeley scholars: concentrated chip supply chain makes a hardwired frontier AI pause feasible — wfithian · 2026-10-16