pub const DECODE_GQA_PACK_HEAD_DIM: u32 = 256;Expand description
2026-09-25: The head dim the packed kernels are compiled for: both sources
#define PD_HDIM 256, and gqa_pack_shape_ok refuses any other head dim.
pub const DECODE_GQA_PACK_HEAD_DIM: u32 = 256;2026-09-25: The head dim the packed kernels are compiled for: both sources
#define PD_HDIM 256, and gqa_pack_shape_ok refuses any other head dim.