[2/N Dynamic Shapes]: Support dynamic dims for single-dispatch (zero-workspace) graphs - #399
Merged
Merged
Conversation
sjain-stanford
changed the title
sjain-stanford
force-pushed
the
users/sambhav/tensor-dynamic-dims
branch
2 times, most recently
from
May 7, 2026 22:49
4213419 to
235c7db
Compare
sjain-stanford
changed the title
sjain-stanford
force-pushed
the
users/sambhav/tensor-dynamic-dims
branch
2 times, most recently
from
May 8, 2026 00:14
d9953f8 to
3002d37
Compare
sjain-stanford
marked this pull request as ready for review
May 8, 2026 00:16
Member
Author
|
Also, cc: @BrianHarrisonAMD for visibility |
sjain-stanford
force-pushed
the
users/sambhav/tensor-dynamic-dims
branch
from
May 8, 2026 05:17
7f63564 to
852b272
Compare
IanWood1
reviewed
May 8, 2026
sjain-stanford
force-pushed
the
users/sambhav/tensor-dynamic-dims
branch
from
May 8, 2026 17:53
852b272 to
abc7fac
Compare
IanWood1
approved these changes
May 12, 2026
sjain-stanford
force-pushed
the
users/sambhav/tensor-dynamic-dims
branch
from
May 12, 2026 23:48
b05b84e to
1132c73
Compare
Signed-off-by: Sambhav Jain <sambhav@alumni.stanford.edu>
Covers the basic MHA SDPA path with dynamic sequence length metadata and keeps the sample expected-fail only on AMDGPU until compiler/runtime support is ready. Signed-off-by: Sambhav Jain <sambhav@alumni.stanford.edu>
Emit dynamic custom-op dimension placeholders as question marks and reject representative unit extents for dynamic dimensions. Signed-off-by: Sambhav Jain <sambhav@alumni.stanford.edu>
Signed-off-by: Sambhav Jain <sambhav@alumni.stanford.edu>
sjain-stanford
force-pushed
the
users/sambhav/tensor-dynamic-dims
branch
from
May 12, 2026 23:57
1132c73 to
2890730
Compare
sjain-stanford
enabled auto-merge (squash)
May 13, 2026 00:02
sjain-stanford
disabled auto-merge
May 13, 2026 00:50
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
3 participants
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.
This is the second dynamic-shapes slice. It enables Fusilli graphs to keep representative static extents in
TensorAttrwhile emitting annotated dimensions as dynamic (?) in Torch MLIR types, which lets IREE compile and run zero-transient/single-dispatch dynamic-shape graphs.This PR covers:
TensorAttrScope and follow-up:
iree.abi.transients.sizeand require invoking the runtime size function before execution.Co-authored-by: GPT 5.5 codex@openai.com
Generated with Codex