Search results

[Pallas] adjust_block_size_constraints max-propagates per-tensor TPU alignment across all tile-indexed tensors sharing a block_id

Status: Open.

#2842 In pytorch/helion;

· norx1991 opened on Jun 22, 2026

[RFC] Helion Remote Autotuning Service

Status: Open.

#2745 In pytorch/helion;

· mrodden opened on Jun 10, 2026

Same autotuned "static_ranges" config fails across different specialized shapes

Status: Open.

#2705 In pytorch/helion;

· xiaohongchen1991 opened on Jun 5, 2026

Compile-phase subprocess isolation to survive Triton C++ aborts

precompile

Status: Open.

#2531 In pytorch/helion;

· choijon5 opened on May 20, 2026

Support subprocess benchmarking for non-picklable epilogue arguments

autotuning

Status: Open.

#2469 In pytorch/helion;

· choijon5 opened on May 16, 2026

Optimize fixed launch overheads

Status: Open.

#2468 In pytorch/helion;

· jansel opened on May 16, 2026

[Pallas] test_atomic_add_w_tile_attr broken with recent torch_tpu

Status: Open.

#2304 In pytorch/helion;

· cota opened on May 6, 2026

Better support for Larger Grouped Operations in Helion in hl.grid

user support

Status: Open.

#2148 In pytorch/helion;

· melody-lui3 opened on Apr 30, 2026

Kernel perf regression using latest main branch with configs autotuned from v1.0.0

performance

user support

Status: Open.

#2044 In pytorch/helion;

· xiaohongchen1991 opened on Apr 17, 2026

[AMD 350x] Helion Layernorm kernel performance is worse compared to inductor for some shapes in backward pass

rocmAMD related issuesAMD related issues

Status: Open.

#2013 In pytorch/helion;

· MengjiaoZhou opened on Apr 13, 2026

not use TMA for symm mem tensor

CUDA error

Status: Open.

#1846 In pytorch/helion;

· shunting314 opened on Mar 27, 2026

Enable Partial Tile Store/Loading

language support

user support

Status: Open.

#1679 In pytorch/helion;

· mengluy0125 opened on Mar 14, 2026