llama.cpp Release b11873 Adds Initial q2_k Quantization for OpenCL

Release b11873 of llama.cpp adds initial q2_k quantization support for OpenCL, introducing specific performance improvements and kernel optimizations for Adreno GPUs. This update significantly improves the efficiency of running large language models on mobile and edge devices utilizing Qualcomm Adreno graphics hardware. By reducing memory footprints through q2_k quantization via OpenCL, mobile deployment becomes much more viable. The pull request implements new OpenCL features including mul_mv_q2_k_f32, q2_k SOA matrix-vector operations, tiled matrix multiplication, and dp4a gemm/gemv operations tuned specifically for Adreno. It was co-authored with contributions from Qualcomm engineers.