perf(detray): Align transform3 struct to 16 bytes - #6027
Draft
stephenswat wants to merge 1 commit into
Draft
Conversation
Member
Author
Performance summaryHere is a summary of the performance effects of this PR: GraphicalTabular
Important All metrics in this report are given as reciprocal throughput, not as wallclock runtime. Note This is an automated message produced upon the explicit request of a human being. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Contributor
Public API surface diffNo change to the public API surface. ✅ |
Contributor
niermann999
previously approved these changes
Sep 4, 2026
niermann999
left a comment
niermann999
left a comment
Contributor
There was a problem hiding this comment.
I also have been playing around with alignment of these during the vectorization studies and never saw much of a difference...
The transform3 struct is exactly 128 bytes (although it can go down to 48 without issue), but is aligned to 4 byte boundaries. This means that any loads into the struct must assume non-alignment. Aligning the struct to 16 bytes allows the compiler to issue 128 bit loads if it wants, which should improve latency and throughput.
stephenswat
force-pushed
the
detray/perf/align_transform3_to_16_bytes
branch
from
September 7, 2026 11:34
574ba42 to
7a6e7b6
Compare
acts-policybot
Bot
dismissed
niermann999’s stale review
September 7, 2026 11:34
Invalidated by push of 7a6e7b6
Member
Author
Performance summaryHere is a summary of the performance effects of this PR: GraphicalTabular
Important All metrics in this report are given as reciprocal throughput, not as wallclock runtime. Note This is an automated message produced upon the explicit request of a human being. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
niermann999
reviewed
Sep 8, 2026
| /// @param m is the matrix | ||
| /// @param col is the column index | ||
| DETRAY_HOST_DEVICE | ||
| static constexpr array_type<4> column(const matrix44 &m, index_t col) { |
Contributor
There was a problem hiding this comment.
I think it would be better to implement this as specialized block getter? That could benefit more code than just the transform3
Contributor
There was a problem hiding this comment.
See e.g. acts-project/algebra-plugins#140
Also, could this go into a dedicated PR?
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.
The transform3 struct is exactly 128 bytes (although it can go down to 48 without issue), but is aligned to 4 byte boundaries. This means that any loads into the struct must assume non-alignment. Aligning the struct to 16 bytes allows the compiler to issue 128 bit loads if it wants, which should improve latency and throughput.