All issues

[v.3.8.0] Release Tracker
#10907 · atalman opened on Jul 16, 2026
27

[06-fused-attention.py] There maybe a bug in the FP8 attention forward kernel.

#11440

· chengjunlu opened

on Aug 25, 2026

[TritonGPU] ttg.warp_specialize capturing a tensor aborts with report_fatal_error and no diagnostic

#11405

· alepot55 opened

on Aug 23, 2026

[AMD] scf.for with runtime trip count miscompiled on gfx1151

#11378

· neuhaus opened

on Aug 20, 2026

[SM80/A100] 2x perf regression 3.6.0 -> 3.7.1 at previously-optimal num_warps=4 (per-config optimum inverts); 13+ min compiles on large BLOCK sizes

#11373

· onesource2026 opened

on Aug 20, 2026

`desc.scatter()` emits `tile::scatter4` on consumer Blackwell (sm_12x) and fails in ptxas instead of erroring at compile time

#11344

· yashb98 opened

on Aug 18, 2026

AMD compilation crash in TritonAMDGPUPrepareIfCombining

#11327

· O1iveira7 opened

on Aug 16, 2026

[Membar] No barrier between a caller's pending shared access and a callee's first one

#11326

· alepot55 opened

on Aug 16, 2026

TritonAMDGPUBlockPingpong produces invalid TTGIR with tl.dot, dynamic loop, and runtime-selected pointer

#11322

· O1iveira7 opened

on Aug 15, 2026

AMD CanonicalizePointers crashes on tt.pointer_range=32 pointer live across control-flow merge with nested early return.

#11321

· O1iveira7 opened

on Aug 15, 2026

Mark synthesized debug compile units as tile dialect

#11231

· jackalcooper opened

on Aug 9, 2026

Interpreter and GPU disagree: valid kernel silently returns wrong result when compiled

#11153

· O1iveira7 opened

on Aug 3, 2026

BLOCK_SIZE-dependent illegal memory access: crashes at 512/1024, passes at other sizes

#11148

· O1iveira7 opened

on Aug 2, 2026