[Pallas] Route meta output-only tensors to CPU under interpret (#2526)

[Pallas] Use fori_loop if loop bounds includes non-constexpr symbolic…

… values (#1927)

[Helion + torch.compile] Add ref baseline kernel count checks (#1786)

Recreate foreign SymInts in helion's own ShapeEnv (#1671)

Signed-off-by: Yanan Cao <gmagogsfm@gmail.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

[pallas-tpu] Switch float16 to HALF_DTYPE in examples for TPU compati…

…bility (#1648)

Update broken _current_generation reference (#1545)

Co-authored-by: ethche <ethche@gmail.com>

Autotune persistent kernels for multi occupancy (#1307)

Support hl.barrier for mega-kernels (#1151)

Enable tests on py314 (#1306)

Fix invalid Triton code for mixed scalar/block indexing in store oper…

…ations when block dimension has size 1 (#1258)