Fixed issue where op=transform was double-calling transform#1037
Merged
cliffburdick merged 2 commits intomainfrom Aug 15, 2025
Merged
Fixed issue where op=transform was double-calling transform#1037cliffburdick merged 2 commits intomainfrom
cliffburdick merged 2 commits intomainfrom
Conversation
Collaborator
Author
|
/build |
tbensonatl
added a commit
that referenced
this pull request
Aug 14, 2025
Signed-off-by: Thomas Benson <tbenson@nvidia.com>
tbensonatl
approved these changes
Aug 14, 2025
Collaborator
tbensonatl
left a comment
tbensonatl
left a comment
There was a problem hiding this comment.
reshape.h needs to be added, but otherwise this change looks good to me.
Collaborator
Author
|
/build |
cliffburdick
pushed a commit
that referenced
this pull request
Aug 25, 2025
* Add new zipvec operator Add a new zipvec operator that zips multiple input operators into a vectorized operation (e.g., an operator with type float3 when zipping x, y, and z coordinates). Signed-off-by: Thomas Benson <tbenson@nvidia.com> * Update zipvec documentation Signed-off-by: Thomas Benson <tbenson@nvidia.com> * Address part of review comments Use cuda::std namespace for template helpers, update copyright data and static assertion comment Signed-off-by: Thomas Benson <tbenson@nvidia.com> * Remove is_narrowing_conversion helpers and add get_impl helper for zipvec operator() methods. * Remove support for half types * Use scalar loads for vectorized types * Handle the sizeof(T) != alignment_by_type<T>() only in load() * Special-case alignment checks for sizeof(T) != alignment_by_type<T>() * Address remaining review feedback Signed-off-by: Thomas Benson <tbenson@nvidia.com> * Remove use of mtie in zipvec (see PR #1037) Signed-off-by: Thomas Benson <tbenson@nvidia.com> * Add back assignment operator with self_type Signed-off-by: Thomas Benson <tbenson@nvidia.com> * Update ZipVecOp class documentation block Signed-off-by: Thomas Benson <tbenson@nvidia.com> --------- Signed-off-by: Thomas Benson <tbenson@nvidia.com>
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
2 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.
When a statement
op = transformoccurred where op was not a tensor, the transform was being called twice incorrectly leading to poor performance. This simplifies the code and calls it once.