b9189
📦 llama-cppView on GitHub →
🐛 1 fixes🔧 1 symbols
Summary
This release addresses an issue by skipping device enumeration in server router mode to avoid unnecessary CUDA primary context creation. It also provides updated binary distributions for numerous operating systems and hardware architectures.
🐛 Bug Fixes
- Skipped device enumeration in router mode on the server side to prevent the creation of a CUDA primary context.