Expose fused Triton TBE bounds checking controls (#4492) - #4492
Closed
stashuk-olek wants to merge 9 commits into
Closed
Expose fused Triton TBE bounds checking controls (#4492)#4492stashuk-olek wants to merge 9 commits into
stashuk-olek wants to merge 9 commits into
Conversation
Contributor
|
@stashuk-olek has exported this pull request. If you are a Meta employee, you can view the originating Diff in D114286097. |
stashuk-olek
force-pushed
the
export-D114286097
branch
from
August 5, 2026 18:00
a5a1d22 to
d689d09
Compare
stashuk-olek
added a commit
to stashuk-olek/torchrec
that referenced
this pull request
Aug 5, 2026
Summary: Just small convenience - expose in the benchmark to test easier. Reviewed By: axeisghost Differential Revision: D114286097
stashuk-olek
force-pushed
the
export-D114286097
branch
from
August 5, 2026 21:00
d689d09 to
82ba611
Compare
stashuk-olek
added a commit
to stashuk-olek/torchrec
that referenced
this pull request
Aug 5, 2026
Summary: Just small convenience - expose in the benchmark to test easier. Reviewed By: axeisghost Differential Revision: D114286097
stashuk-olek
force-pushed
the
export-D114286097
branch
from
August 6, 2026 16:19
82ba611 to
82ca9dd
Compare
stashuk-olek
added a commit
to stashuk-olek/torchrec
that referenced
this pull request
Aug 6, 2026
Summary: Just small convenience - expose in the benchmark to test easier. Reviewed By: axeisghost Differential Revision: D114286097
stashuk-olek
force-pushed
the
export-D114286097
branch
from
August 6, 2026 20:10
82ca9dd to
454d1ff
Compare
stashuk-olek
added a commit
to stashuk-olek/torchrec
that referenced
this pull request
Aug 6, 2026
Summary: Just small convenience - expose in the benchmark to test easier. Reviewed By: axeisghost Differential Revision: D114286097
stashuk-olek
added a commit
to stashuk-olek/torchrec
that referenced
this pull request
Aug 6, 2026
Summary: Just small convenience - expose in the benchmark to test easier. Reviewed By: axeisghost Differential Revision: D114286097
stashuk-olek
force-pushed
the
export-D114286097
branch
from
August 6, 2026 20:11
454d1ff to
044adf1
Compare
stashuk-olek
added a commit
to stashuk-olek/torchrec
that referenced
this pull request
Aug 6, 2026
Summary: Just small convenience - expose in the benchmark to test easier. Reviewed By: axeisghost Differential Revision: D114286097
stashuk-olek
force-pushed
the
export-D114286097
branch
2 times, most recently
from
August 6, 2026 23:04
03234ec to
f6ab890
Compare
stashuk-olek
added a commit
to stashuk-olek/torchrec
that referenced
this pull request
Aug 6, 2026
Summary: Just small convenience - expose in the benchmark to test easier. Reviewed By: axeisghost Differential Revision: D114286097
stashuk-olek
force-pushed
the
export-D114286097
branch
from
August 7, 2026 17:42
f6ab890 to
ae4cb0f
Compare
Summary: Use BoundsCheckMode.V2 for Triton TBE standalone index validation. This preserves the existing warning and offset-repair behavior while reducing validation overhead before Triton forward. Reviewed By: axeisghost, TroyGarden Differential Revision: D114279989
Summary: Validate and repair indices inside supported unweighted, non-VBE Triton forward kernels. Unsupported configurations retain the CUDA v2 fallback, while fused checked loads preserve warning accumulation without a separate full-index pass. Reviewed By: TroyGarden, axeisghost Differential Revision: D114284802
Summary: Restructure short- and long-run gradient accumulation to reduce serialized reduction pressure. The new reduction schedule exposes more independent work while preserving the existing exact rowwise optimizer update. Reviewed By: axeisghost Differential Revision: D114273026
) Summary: Reduce the dout reduction tile from 16 to 8 in the unweighted short-run, separate long-run accumulation, and fused long-run kernels. The smaller tile reduces live vector state and register pressure on B200 while retaining FP32 accumulation and the existing rowwise optimizer update. Weighted kernels remain unchanged. Reviewed By: axeisghost Differential Revision: D114403703
Summary: Process multiple bags per Triton program where the workload has enough parallelism. This amortizes program overhead and exposes independent loads without changing the embedding reduction. Reviewed By: axeisghost Differential Revision: D114280197
Summary: Issue up to eight independent index gathers per loop iteration on profitable long-bag shapes. The wider loop increases memory-level parallelism while retaining the four-wide path for shapes where register pressure dominates. Reviewed By: axeisghost Differential Revision: D114280947
Summary: Reuse exact counts from an eligible leading, unique small-table feature during exact rowwise Adagrad backward. Immutable plans carry actual table geometry, compensated FP16 high/residual matrix products preserve FP32 accuracy, and scalar validity enables safe fallback for long bags, aliases, graph capture, weighted inputs, and VBE. Reviewed By: axeisghost Differential Revision: D114284800
Summary: Thread the fused-bounds-check option through TorchRec fused parameters and benchmark configuration. The default remains unchanged, while supported callers can select the fused Triton route without bypassing validation. Reviewed By: axeisghost Differential Revision: D114286097
stashuk-olek
force-pushed
the
export-D114286097
branch
from
August 8, 2026 06:14
ae4cb0f to
5f11e1e
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary:
Thread the fused-bounds-check option through TorchRec fused parameters and benchmark configuration. The default remains unchanged, while supported callers can select the fused Triton route without bypassing validation.
Reviewed By: axeisghost
Differential Revision: D114286097