Skip to content

Removing skips and in some cases adding skipIfNotCUDA for cuda only features.#1766

Closed
umechand-amd wants to merge 1203 commits intopytorch:mainfrom
umechand-amd:umechand/remove_skips
Closed

Removing skips and in some cases adding skipIfNotCUDA for cuda only features.#1766
umechand-amd wants to merge 1203 commits intopytorch:mainfrom
umechand-amd:umechand/remove_skips

Conversation

@umechand-amd
Copy link
Copy Markdown
Collaborator

@umechand-amd umechand-amd commented Mar 20, 2026 •
edited
Loading

This PR un-skips tests on ROCM which passed on both MI325 and Mi350 and adds correct skip conditions based on memory size and features which are Cuda-only.

karthickai and others added 30 commits February 10, 2026 03:23
@karthickai
Fix hl.rand/hl.randint crash when dim is specialized to size 1 (pytor…
2de1c80 
…ch#1401)
@yvonne-lab
Update num_warps and num_stages range for AMD (pytorch#1368)
c39e795
@yf225
[CI] Fix B200 flaky test errors by deferring CUDA initialization to t…
5587e25 
…est execution time for pytest-xdist compatibility (pytorch#1399)
@drisspg
Add generic kernel config adapter (pytorch#1385)
3814ef7
@yf225
Disable autocast during kernel compilation to fix mixed-precision tra…
d1eb3b0 
…ining (pytorch#1402)
@oulgen
Add nvfp4 gemm (pytorch#1403)
e00e53b
@yf225
Fix unpacking of torch.sort/topk return types in Helion kernel (pytor…
72b4eb6 
…ch#1406)
@yf225
Skip O(N^2) argsort in torch.sort lowering when indices is unused (…
d1c4ac7 
…pytorch#1408)
@yf225
[CI] Adjust atol for squeeze_and_excitation_net_bwd unit tests (pyt…
94a3771 
…orch#1409)
@yf225
[CI] Adjust atol for test_dtype_cast_preserved_before_second_dot te…
4908b15 
…st (pytorch#1411)
@yf225
Fix torch.cumsum overwriting its input when used in Helion kernel (p…
3a42754 
…ytorch#1410)
@karthickai
Skip hl.rand/hl.randint specialize test on MTIA (pytorch#1414)
b0336f3
@yf225
Skip "too many blocks in cooperative launch" error during autotuning (p…
85e93fe 
…ytorch#1417)
@yf225
[CI] Fix flaky OOM error for test_int32_offset_out_of_range_error o…
5ccd00b 
…n H100 (pytorch#1419)
@yf225
[CI] Adjust atol for test_hl_arange_non_power_of_2 test (pytorch#1421)
31aa4db
@yf225
[Helion + torch.compile] Add multi-output and scalar return support (p…
da94f03 
…ytorch#1395)
@yf225
Improve DeviceIR.get_root() to handle multiple reduction dimensions (…
3b5eaf1 
…pytorch#1434)
@oulgen
Introduce Backend abstraction for multi-backend code generation (pyto…
b2aa111 
…rch#1426)
@oulgen
Route codegen decorator, constexpr, and imports through Backend (pyto…
7b62c12 
…rch#1427)
@oulgen
Route dtype string conversions through Backend abstraction (pytorch#1428
16a359d 
)
@oulgen
Rename triton_index_type to backend-agnostic index_type (pytorch#1429)
958e756
@v0i0
dump triton failures (pytorch#1407)
4a3dd4b
@yf225
Fix PointwiseLowering rank mismatch for size-1 reductions (pytorch#1424)
5705332
@yf225
[Helion + torch.compile] Fix aliasing, mutation detection, and nested…
5214a5b 
… return support (pytorch#1432)
@v0i0
for now rewrite gdn fwd h without hl.grid for fusion (pytorch#1391)
efcca42
@oulgen
Add PallasBackend class and make backend configurable via Settings (p…
929228d 
…ytorch#1447)
@shunting314
sync compute_repeat output for distributed kernels (pytorch#1435)
938bb06
@shunting314
raise tolerance for test_hl_dot_out_dtype_argument (pytorch#1457)
3e5e342
@oulgen
Skip too many resource error in autotuner (pytorch#1458)
a2dbfbe
@gmagogsfm @claude
Preserve symbolic shapes in HOP output_spec during tracing (pytorch#1454
4d11e5b 
)

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
nullplay and others added 3 commits March 20, 2026 11:07
@nullplay
Add hl.jagged_tile (pytorch#1651)
c8ba62d
@bringlein @ibm-bob @claude
Fix logging to be compatible with pytest (pytorch#1734)
f83f6c2 
Co-authored-by: AskBob <askbob@ibm.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
@bringlein @ibm-bob @claude
Add autotune_initial_population_strategy kernel setting (pytorch#1735)
ae6ab3a 
Co-authored-by: AskBob <askbob@ibm.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
@umechand-amd umechand-amd force-pushed the umechand/remove_skips branch from 34a7e37 to aadb23b Compare March 20, 2026 18:54
@umechand-amd umechand-amd force-pushed the umechand/remove_skips branch from aadb23b to 427a73f Compare March 20, 2026 23:19
aditvenk and others added 11 commits March 20, 2026 21:05
@aditvenk
Enable pyrefly on macOS with ignore-missing-imports (pytorch#1760)
761b131
@aditvenk
[metal] Register "metal" backend with minimal MetalBackend and launch…
be7effe 
…er (pytorch#1761)
@aditvenk
[metal] Respect force_tile_mask() in NDTileStrategy mask generation (p…
45b18fe 
…ytorch#1762)
@choijon5
Remove past hackathon event from README (pytorch#1780)
036103f
@choijon5
Remove deadcode _clone_tree and _assert_args_close (pytorch#1781)
ae67384
@choijon5
Fix PermutationFragment.encode() returning wrong value (pytorch#1779)
5134f91
@choijon5
Fix atomic_max ref to return previous value (pytorch#1782)
0ff9601
@jansel
Fix lints (pytorch#1775)
1346126
@jansel
Add runpod SKILL.md (pytorch#1776)
6dbff03
@yf225
[Helion + torch.compile] Add store/load transform hooks and prologue/…
76067b2 
…epilogue fusion codegen (pytorch#1724)
@yf225
[Helion + torch.compile] Enable torch.compile fusion tests (pytorch#1727
33b4dcb 
)
@umechand-amd umechand-amd force-pushed the umechand/remove_skips branch from f10ff78 to 89804d5 Compare March 23, 2026 18:23
@umechand-amd umechand-amd marked this pull request as ready for review March 23, 2026 18:29
@umechand-amd umechand-amd force-pushed the umechand/remove_skips branch from b9cf440 to 5344848 Compare March 23, 2026 21:00
@umechand-amd umechand-amd deleted the umechand/remove_skips branch March 25, 2026 21:06
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

CLA Signed This label is managed by the Meta Open Source bot.

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.