ggml/llama.cpp releases: b10867

TL;DR

The ggml/llama.cpp project updated its codebase to disable lazy tensor loading by default on integrated GPUs (iGPUs) to address performance issues, impacting users of these less powerful graphics cards. This change is crucial for ensuring smoother model operation across different hardware configurations without sacrificing functionality.

Detailed Summary

The ggml/llama.cpp project released a new update disabling lazy tensor loading by default on integrated GPUs (iGPUs) to address performance issues, while also adding an automatic lazy mode that can be toggled. This change impacts users of the software who experience regression on iGPUs and could improve compatibility across different hardware configurations. The broader impact includes enhanced usability for a wider range of computing devices, particularly those with limited GPU resources.

Key Points

• • llama: disable lazy tensor loading by default on iGPUs (#28326)
• • llama: add lazy mode auto, fix iGPU regression
• • revert changes except disabling lazy load on iGPUs in AUTO