Tracks model size at load time and checks Metal active memory after
each generation. If usage exceeds 3× model size, forces double GC
and cache clear as a safety net.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Go GC cannot see Metal/C memory pressure, so intermediate arrays from
each forward pass accumulated without bound, causing OOM kills after
3-4 requests. Fix: runtime.SetFinalizer on every Array releases C
handles when GC collects them, and runtime.GC() is forced every 4
tokens during generation. Also adds SetMemoryLimit(24GB) as a hard
Metal ceiling.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add ClearCache() wrapping mlx_clear_cache
- Clear Metal allocator cache every 8 tokens during generation
- Set 16GB cache limit on backend init
- Prevents GPU memory from growing unbounded during inference
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
CGo wrapper for mlx-c providing zero-Python Metal GPU inference.
Includes Gemma 3 model architecture, BPE tokenizer, KV cache,
composable sampling, and OpenAI-compatible serve command.
Build-tagged (darwin && arm64 && mlx) with stubs for cross-platform.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
