ggml/llama.cpp releases: b10909

TL;DR

The ggml/llama.cpp project updated its Metal backend to consolidate all fusable operation patterns into a single fusion table, enhancing debug capabilities. This change aims to improve optimization and maintainability of the codebase.

Detailed Summary

The ggml/llama.cpp project has released updates including a rework of fusion patterns into a single table for the Metal backend, consolidating all fusable operation patterns. This change impacts the graph optimization process and debug capabilities. Broader implications include improved efficiency and easier maintenance of the codebase.

Key Points

• • metal : rework fusion patterns into a single table
• • All fusable op patterns for Metal backend are consolidated
• • Fusion debug rework included in the update