Mac Metal GPU Access Fails Over SSH, Fix Requires GUI Domain Relaunch
arthurcolle · x · 2026-08-16
Running llama-server via plain SSH on a Mac resulted in slow inference (5 tok/s) despite the Metal dylib being linked.
Root Cause:
When a process is launched over SSH without a GUI session context, Metal cannot detect a GPU device, forcing a fallback to CPU-only mode (logs showed nthreads=8 and zero Metal init). LM Studio works because it runs inside the logged-in GUI session.
Fix:
Relaunch the llama-server in the GUI domain using launchctl to ensure it inherits the necessary Metal context and GPU acceleration.
More from Infra
- CoreWeave Growth Outpaces AWS: AI Cloud Market Dynamics — abhiadesai · 2026-08-16
- DDR4 Memory Is a Cash Cow for Nanya: Legacy Memory Outperforms Leading Edge — zephyr_z9 · 2026-08-16
- Data center bans are illegitimate; regulate harms, not existence — QuintinPope5 · 2026-08-16
- Prefer buying physical hardware and building a data center over subscriptions — arthurcolle · 2026-08-16
- Does RX 6700XT Get Official ROCm Support? User Reports ComfyUI Working — Aromatic-Lie-7056 · 2026-08-16
- Helium Browser's New Feature Splits TLS ClientHello to Bypass Censorship, Seeking Feedback — uwukko · 2026-08-16