NVIDIA/TileGym v1.4.0
NVIDIA/TileGym
Captured source
source ↗published Jul 25, 2026seen 1dcaptured 1dhttp 200method plain
v1.4.0
Repository: NVIDIA/TileGym
Tag: v1.4.0
Published: 2026-07-25T03:07:57Z
Prerelease: no
Release notes:
What's Changed
- cutile recurrent_gated_delta_rule: autotune + persistent variant & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/129
- [recurrent_gated_delta_rule] Add stream-v kernel and tuning huristics & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/131
- ci: add request-nvskills-ci.yml workflow by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/133
- Add tilecpp backend by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/134
- feat(softmax): add explicit multi-wave kernel option by @liqiangxl in https://github.com/NVIDIA/TileGym/pull/128
- Initial NVSkills-CI onboarding for TileGym skills by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/135
- benchmark_fn_cupti: surface kernel_name and kernel_times in result dict & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/138
- Skill Evals: tilegym-cutile-python by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/144
- Add pre-commit CI checks and skills-signature validation by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/137
- Add FlashInfer-style transformer kernel inventory by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/139
- Fix pre-commit format failure and update format.sh by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/149
- Skill Evals: cutile-to-triton + improve-cutile-perf + cutile-to-julia by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/152
- add some cuTile kernels for Liger suites & Update mla autotune config & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/151
- perf(liger/cutile): tune cross_entropy BLOCK_SIZE by GPU arch & Modularize transformer HF benchmark & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/153
- [perf][cutile]autotune RoPE occupancy & Fix flashinfer cutile kernels & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/156
- Update CI image to CUDA 13.3 / Ubuntu 24.04 and shard the GPU test & benchmark jobs by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/158
- ci: refresh pytest-split .test_durations by @github-actions[bot] in https://github.com/NVIDIA/TileGym/pull/159
- Update recurrent_gated_delta_rule and matmul & Add FlashInfer-backed transformer kernel inventory workflow by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/160
- ci: add transformer inventory runtime test to GitHub CI by @lirundong in https://github.com/NVIDIA/TileGym/pull/162
- [cutile] matmul: expose ct.load cost (latency) as an autotune knob & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/161
- fix result mismatch for Test_GroupGemm & Add Triton backend support for ops kernels & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/163
- ci: refresh pytest-split .test_durations by @github-actions[bot] in https://github.com/NVIDIA/TileGym/pull/165
- [tests]Shrink matmul 16384^3 config to 8192^3 to avoid SM89 OOM & Update tilecpp tests & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/164
- ci: refresh pytest-split .test_durations by @github-actions[bot] in https://github.com/NVIDIA/TileGym/pull/167
- Update matmul op args & Update test_mla_decoding_split_kv parameter & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/166
- Raise cuda.tile's autotune launch-timeout budget when requested via env & other updates by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/168
- ci: refresh pytest-split .test_durations by @github-actions[bot] in https://github.com/NVIDIA/TileGym/pull/170
- Bump version from 1.3.0 to 1.4.0 by @hannahli-nv in https://github.com/NVIDIA/TileGym/pull/172
New Contributors
- @github-actions[bot] made their first contribution in https://github.com/NVIDIA/TileGym/pull/159
- @lirundong made their first contribution in https://github.com/NVIDIA/TileGym/pull/162
Full Changelog: https://github.com/NVIDIA/TileGym/compare/v1.3.0...v1.4.0