Function gemm_blk_launch_threads 

pub fn gemm_blk_launch_threads(f: u32, bm: usize, k: usize) -> usize

The threads one launch of the block-format tensor-core GEMM over k takes: a Q8_0 matrix’s streamed path over 128 rows, eight warps of 64 rows by 32 outputs — two blocks of them, their registers and stages, share a multiprocessor — and a warp each 32 × 32 of the tile otherwise (gemm_blk_threads).