Make stats tests numerically more stable for oneAPI #662
Conversation
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
Contributor
|
Your PR no longer requires formatting changes. Thank you for your contribution! |
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Outdated suggestions cannot be applied.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.
On oneAPI.jl, the cor(A; dims=2) test failed for
Float16(one some GPUs), caused by numerical instability when computing the variance of rows with only 2 elements drawn from [0, 1). This leads to subnormal variances, which are flushed to zero on the CPU (causing NaN or -1 results) but preserved on the oneAPI GPU (causing valid results), resulting in a mismatch.Now I'm not a stats person, and I don't want to fix anything deeper in Intel oneAPI, but it seems to be fixed with an extended range. I saw that there was already such a special case for
Real. Unfortunately, that still doesn't fix theComplexF32case.