2 Commits

Author	SHA1	Message	Date
biondizzle	bd16e8fa85	fix: use tcgen05.wait::st/ld instead of nonexistent tcgen05.fence ... ROOT CAUSE of TMET hang: tcgen05.fence.cta_group::1.sync.aligned is NOT a valid PTX instruction. The correct TMEM ordering primitives are: - tcgen05.wait::st.sync.aligned (wait for TMEM stores to complete) - tcgen05.wait::ld.sync.aligned (wait for TMEM loads to complete) Found in cutlass/arch/barrier.h fence_view_async_tmem_store/load.	2026-05-28 07:12:26 +00:00
biondizzle	ba1e81f2dc	test: minimal TMEM isolation test (alloc, store, load, dealloc)	2026-05-28 07:09:06 +00:00