local-llm
- The deadlock test can't see a half-held lock
Every model wrote a thread test. Most of them could not fail: with locks taken in id order, opposite-direction transfers serialize on one mutex.
- Zig's defer ends at the closing brace
A local model locked two accounts in the right order, scored 97/100, and lost money. One Zig defer sat inside an if block and unlocked at the brace.
- I blamed OpenCode; a 2-bit model passed Zig in it
Four 27B models failed a Zig agent task in OpenCode, and the client looked guilty. Then a 2-bit MoE in the same setup passed 6 of 6.
- 128k context is the floor, and prefill is the wall
A 128k context fits easily in 128 GB of unified memory. Filling it took gpt-oss-120b 29 minutes, and one model's 64k prefill hit a GPU ring timeout.
- The download that contaminated the benchmark
A model download bent one row of a long-context benchmark. The queue built to stop that got deleted by systemd at boot, then met a real power cut.
- 120 GiB for the GPU, set in the kernel, not the BIOS
On a 128 GB Strix Halo box the BIOS's 96 G GPU setting leaves Linux 31 GiB. The minimum plus two TTM parameters gives the GPU 120 GiB.