Ollama: Weekly Recap - MLX Performance & Path Handling
This week brought major MLX cache optimizations and comprehensive fixes to file path handling across the CLI. The team also expanded hardware support with new backends and resolved critical GPU detection issues on hybrid systems.
Duration: PT3M
Episode overview
This episode is a short developer briefing from Ollama.
It explains recent repository work in plain language.
- Show: Ollama
- Published: 2026-06-15T09:08:58Z
- Audio duration: PT3M
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Good morning. This is your Ollama weekly recap for June 8th through 15th, 2026.
Fifty pull request activity items and thirteen additional commits this week focused heavily on performance optimization and user experience polish.
The strongest pattern this week was a comprehensive overhaul of MLX performance infrastructure. Jesse Gross merged PR 16363, introducing significant cache improvements that consolidate MTP speculation onto existing snapshot mechanisms. This reduces complexity while enabling speculation with recurrent caches, setting…
File path handling saw extensive fixes across the CLI interface. Multiple PRs addressed a persistent issue where paths containing directories with file extensions were incorrectly parsed. PR 16735 and several related submissions ensure that paths like "directory.png/image.png" are now treated as complete file paths…
Hardware compatibility expanded significantly this week. PR 16736 adds ZenDNN backend support for AMD EPYC CPUs, while PR 16720 fixes GPU memory detection on unified-memory APUs like AMD's Strix Halo series. The Windows hybrid graphics fix in PR 16669 corrects inverted integrated and discrete GPU classification that…
Several…
Nearby episodes from Ollama
- Gemma 4 Support and Platform Improvements
- Memory Management and Multimodal Parsing Fixes
- GPU Offloading and Tool Call Fixes
- Performance Optimizations and Model Handling Improvements
- Infrastructure Updates and Platform Fixes
- Multimodal Fixes and Developer Experience Updates
- Cache Architecture Overhaul and Data Race Fixes
- Developer Tools and Cross-Platform Reliability