llama.cpp b9788 release enhances dual-GPU support

How it developed

1. June 26, 2026llama.cpp ReleasesWhere it started

   llama.cpp b9788 release enhances dual-GPU support

2. June 28, 2026llama.cpp Releases

   Llama.cpp Enhances OpenCL Flash Attention

   Llama.cpp adds OpenCL flash attention support for f16 and f32 types, optimizing memory usage and adding q4_0 and q8_0 data type support.

3. June 28, 2026llama.cpp Releases

   llama.cpp b9827 release enhances CUDA performance