Gfx11#29

Draft

liangliangchang wants to merge 13 commits into

liangliangchang:gfx11

Commits on Jun 23, 2026

Add kernel dimension logging for profiling analysis

liangliangchang
and
claude
committed
Fix M/N/K dimension logging for MMQ and MMVQ profiling.

liangliangchang
and
cursoragent
committed
Improve RDNA3.5 MMQ occupancy by reducing mmq_x when LDS limits WGs/CU.

liangliangchang
and
cursoragent
committed
Improve RDNA3.5 MMQ prefill via LDS layout and activation load pipelining.

liangliangchang
and
cursoragent
committed
Add opt-in HIP MMQ phase profiling to split dequant vs MMA time.

liangliangchang
and
cursoragent
committed
Improve RDNA3.5 Q4_K MMQ dequant/vec_dot and fix phase profiling overhead.

liangliangchang
and
cursoragent
committed

Commits on Jun 25, 2026