Xuyang LiandClaude Opus 4.8 3953e4b5fc [BugFix] Reject T.alloc_barrier() on pre-Hopper targets with a clear error (#2345) ...

LowerSharedBarrier lowers T.alloc_barrier() to hardware mbarrier code (the
cutlass Barrier type, tl::tl_shuffle_elect, tl::fence_barrier_init), all of
which only exist on sm_90+. On pre-Hopper targets (e.g. sm_86) the frontend
accepted the call silently and nvcc later failed with cryptic
"identifier 'Barrier' is undefined" / "namespace 'tl' has no member ..."
errors.

Add a static compute-capability guard in CUDAPassPipelineBody, before
LowerSharedBarrier: if the target is < sm_90 and the module uses a shared
barrier, raise a clear ValueError naming the offending primitives and pointing
to __syncthreads()/named barriers as the pre-Hopper alternative.

Fixes the sm_86 alloc_barrier compile crash.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

2026-06-07 23:07:22 +08:00

..

test_tilelang_alloc_barrier_target_guard.py

[BugFix] Reject T.alloc_barrier() on pre-Hopper targets with a clear error (#2345)

2026-06-07 23:07:22 +08:00

test_tilelang_codegen_cutedsl_cp_async.py

[CuTeDSL] Integrate host codegen call sites (#2221)

2026-05-21 20:28:41 +08:00

test_tilelang_codegen_cutedsl_scan.py

[TileOP] Add scan operators (#2262)

2026-05-25 17:41:10 +08:00

test_tilelang_rocm_target.py

[AMD][RDNA4] Fix gfx12 (RDNA 4 / Wave32) related CI issues (#2313)

2026-06-03 11:20:30 +08:00