KEMBAR78
Fixes a number of bugs, todos, etc. by schweitzpgi · Pull Request #3083 · NVIDIA/cuda-quantum · GitHub
Skip to content

Conversation

@schweitzpgi
Copy link
Collaborator

@schweitzpgi schweitzpgi commented Jun 18, 2025

  • Reorder two QIR codegen passes to avoid silly inefficiency.
  • Fix a bug for when the lookup fails.
  • Add guard to avoid crash.
  • Reorganize apply specialization to prop constants before generating control, adjoint, both functions rather than after.
  • See also Enable the SROA pass #3064.

- Reorder two QIR codegen passes to avoid silly inefficiency.
- Fix a bug for when the lookup fails.
- Add guard to avoid crash.
- Reorganize apply specialization to prop constants before
  generating control, adjoint, both functions rather than
  after.
- Make the filter less aggressive.
- Add CG dialect.
- Change the patterns for use with greedy.

Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>
@schweitzpgi schweitzpgi requested review from 1tnguyen, annagrin, bmhowe23 and khalatepradnya and removed request for bmhowe23 June 25, 2025 23:42
@schweitzpgi schweitzpgi enabled auto-merge (squash) June 25, 2025 23:43
Copy link
Collaborator

@1tnguyen 1tnguyen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM 👍

Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>
profile. Assuming these should be folded. However, they are now
turning up in python/tests/mlir/target/{graph_coloring,qspan_slices}.py

Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>
Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>
Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>
TODO: mapping needs to be redone to be composable.

Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>
@schweitzpgi schweitzpgi disabled auto-merge June 26, 2025 16:40
@schweitzpgi schweitzpgi enabled auto-merge (squash) June 26, 2025 16:42
Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>
@schweitzpgi schweitzpgi merged commit e247264 into NVIDIA:main Jun 26, 2025
195 checks passed
github-actions bot pushed a commit that referenced this pull request Jun 26, 2025
@schweitzpgi schweitzpgi deleted the ch-grover.6 branch June 26, 2025 19:42
@bettinaheim bettinaheim changed the title [core] Fixes a number of bugs, TODOs, etc. Fixes a number of bugs, todos, etc. Jul 22, 2025
@bettinaheim bettinaheim added the bug fix To be listed under Bug Fixes in the release notes label Jul 22, 2025
@bettinaheim bettinaheim added this to the release 0.12.0 milestone Jul 22, 2025
atgeller pushed a commit to atgeller/cuda-quantum that referenced this pull request Jul 31, 2025
* [core] Fixes a number of bugs, TODOs, etc.

- Reorder two QIR codegen passes to avoid silly inefficiency.
- Fix a bug for when the lookup fails.
- Add guard to avoid crash.
- Reorganize apply specialization to prop constants before
  generating control, adjoint, both functions rather than
  after.
- Make the filter less aggressive.
- Add CG dialect.
- Change the patterns for use with greedy.

* Eliminate variable per code review.
* Add `private` for filtering out unused functions.
* Fix mapping to deal with multiple functions being present in a module.

TODO: mapping needs to be redone to be composable.

---------

Signed-off-by: Eric Schweitz <eschweitz@nvidia.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug fix To be listed under Bug Fixes in the release notes

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants