hex-mm

This release introduces significant enhancements to the Hexagon backend, including multi-NPU device support, a fully asynchronous backend, and improved performance for various operations like ALLREDUCE and matrix multiplications. It also includes numerous bug fixes and optimizations across different Hexagon modules.

This release introduces a significant rework of the L2 cache handling in the hexagon backend, focusing on dirty bit tracking and lazy flushing for improved performance. It also includes numerous updates and fixes across various hexagon and HMX components, enhancing thread management, DMA operations, and overall robustness.

This release focuses heavily on reworking the hexagon flash attention implementation, bringing significant optimizations and accuracy improvements across various internal components (hex-mm, hex-fa, hmx-fa). Numerous bug fixes and performance enhancements related to tracing, memory alignment, and kernel usage were also implemented.

This release introduces a major rework of hexagon matrix multiplication (MUL_MAT/MUL_MAT_ID) with new tiled weight repacking and performance optimizations across HVX and HMX backends. Support for hardware older than architecture v73 has been removed.

This release focuses heavily on cleanup and performance optimizations across Hexagon (HEX) and HMX backends for matrix multiplication (MUL_MAT, MUL_MAT_ID), Flash Attention, and GDN, introducing initial F32 matmul support and fixing several fusion and stride bugs.

This release introduces significant performance enhancements for Hexagon devices by adding Q4_1 quantization support for matrix operations. Several related bugs in buffer sizing and fp16 calculations were also resolved.

This release includes internal optimizations for Hexagon processing, specifically moving M-tail row processing to HMX and optimizing the padded activation loop in hmx-mm.

This release updates the Hexagon module by bumping the HMX frequency to its maximum corner setting and resolves a logging error in hex-mm.

This release significantly improves hexagon performance through op request batching, buffer management rewrite, and explicit L2 cache control. It also removes the deprecated GGML_HEXAGON_EXPERIMENTAL environment variable.

This release focuses on Q4_0 and MXFP4 repack fixes within the hexagon backend, optimizing matrix multiplication kernels and resolving tail corruption issues.

This release focuses heavily on Flash Attention optimizations for hexagon, including DMA, mpyacc, and multi-row improvements, alongside various MatMul updates and bug fixes in hexagon vector operations.