b11632: cuda, hip: GDN chunked kernel (#29353)

<ul> <li> <p>CUDA: faster MMVQ for low-bit types</p> </li> <li> <p>ggml-cuda: hip: small nwarps GCN fix on Q3_K(ncols=2,3), Q6_K(ncols=4)</p> </li> </ul> <hr /> <p>Co-authored-by:

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1x5skta/cuda_hip_gdn_chunked_kernel_by_am17an_pull/"> <img alt="cuda, hip: GDN chunked kernel by am17an ·

<ul> <li> <p>opencl: add bin kernel <code>gemm_moe_mxfp4_q8_1_dp4a_ila_ldck</code> for X1</p> </li> <li> <p>opencl: fix whitespace</p> </li> <li> <p>opencl: undo unnecessary

<ul> <li>cuda: cap Stream-K blocks at the number of k-blocks</li> </ul> <p>Stream-K normally launches one CUDA block per SM. For small matrices,<br /> this can exceed the total

<ul> <li>cuda: stage the lightning indexer queries in head passes for MUSA</li> </ul> <p>MUSA archs 21 and 22 cap static shared memory at 28 KB, and the tile<br /> kernel staged

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1svdjfa/cuda_reduce_mmq_streamk_overhead_by/"> <img alt="CUDA: reduce MMQ stream-k overhead by

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1s71d3f/optimize_moe_gemv_kernel_for_bs_1_by_gaugargnv/"> <img alt="Optimize MOE GEMV kernel for BS &gt; 1.

<ul> <li> <p>CUDA: refactor swizzling code</p> </li> <li> <p>fix templates/loop bounds</p> </li> </ul>