Some precision improvements for serial kernels #518
Closed
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Outdated suggestions cannot be applied.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 contains several fixes that improve the precision in some serial kernels to match those in the normal implementation.
abs2
to calculateaf
do not re-calculateaf
byfdev + mag
, but storing itThe operations above are mathematically the same before and after the changes, but associative property does not generally hold for floating-point arithmetic, so the above changes ensure the implementation in serial kernels can get the same precision with the normal implementation.
EDIT: Memory is probably more important than precision here, so we don't use additional storage for
af
.