Fix FPRF handling of denormal singles #9811
Merged
+224
−194
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
JosJuice
force-pushed
the
fprf-denormal-singles
branch
2 times, most recently
from
286fdbb
to
08ef45b
Compare
Sintendo
reviewed
Jun 13, 2021
JosJuice
force-pushed
the
fprf-denormal-singles
branch
from
08ef45b
to
61720cf
Compare
The interpreter was wrong after all. Hardware verified.
Performance optimization, along with making the code a little neater. Saves us from performing a single -> double -> single conversion when calling UpdateFPRFSingle.
JosJuice
force-pushed
the
fprf-denormal-singles
branch
from
61720cf
to
ccd8233
Compare
Tilka
approved these changes
Jun 28, 2021
phire
approved these changes
Jun 28, 2021
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Outdated suggestions cannot be applied.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.
We were classifying denormal singles as if they were normals (because we converted singles to doubles and then classified them using the same code as for doubles), but this behavior has been confirmed incorrect by dolphin-emu/hwtests#41.
When fixing this, I went with an approach of performing classification on 32-bit floats in all CPU emulation cores (i.e. classification is performed after forcing the result to single but before converting it back to double). It would be possible to perform classification after converting to double, but that would require us to check if the exponent is in a certain range instead of just checking if it is zero, so I think the approach I went with is better.