Improve SoA with static array extents #653
Merged
Conversation
This file contains 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
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## develop #653 +/- ##
========================================
Coverage 98.75% 98.75%
========================================
Files 74 74
Lines 6818 6828 +10
========================================
+ Hits 6733 6743 +10
Misses 85 85 |
bernhardmgruber
force-pushed
the
soastaticoffsets
branch
2 times, most recently
from
d13540a
to
ae82cba
Compare
bernhardmgruber
changed the title
|
Runtime difference is marginal:
compile time: runtime:
|
|
Here, the CT variant is actually a bit slower. But the assembly is better (less instructions). |
bernhardmgruber
force-pushed
the
soastaticoffsets
branch
4 times, most recently
from
97902ee
to
a4b95c1
Compare
bernhardmgruber
force-pushed
the
soastaticoffsets
branch
2 times, most recently
from
aa8902d
to
88cc497
Compare
bernhardmgruber
force-pushed
the
soastaticoffsets
branch
2 times, most recently
from
c5d0774
to
e3c32eb
Compare
bernhardmgruber
force-pushed
the
soastaticoffsets
branch
from
a31a94c
to
ca64099
Compare
* Mix data types in record dimension, so padding and alignment can have an effect * Make problem size an odd number * Plot results * Test all LLAMA versions in a single run * Test a few variations of the SoA layout * Simplify alpaka use
This works around an MSVC bug.
bernhardmgruber
force-pushed
the
soastaticoffsets
branch
from
ca64099
to
1ee97e4
Compare
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Outdated suggestions cannot be applied.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.
This PR improves the SoA by precomputing a compile-time only subarray offset table when the array extents are fully static.
This PR also contains an extension of the vectoradd benchmarks:
* Mix data types in record dimension, so padding and alignment can have an effect
* Make problem size an odd number
* Plot results
* Test all LLAMA versions in a single run
* Test a few variations of the SoA layout
* Simplify alpaka use