Improve fmath.h madd function #2492
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
I was previously trying to use std::fma, but upon further investigation it was turning into a single vfmadd instruction a lot less than I thought, in many cases ending up as a function call (which itself was extra slow when emulating fma on hardware that lacked it). The better strategy seems to be just saying `a*b+c`, which on gcc & icc compilers automatically turns into vfmadd when available on the hardware, and adding a clang-specific pragma ensures this behavior for clang as well. Thanks to Alex Wells of Intel for pointing this out to me (in OSL land).
lgritz
added a commit
to lgritz/OpenImageIO
that referenced
this pull request
Feb 27, 2020
I was previously trying to use std::fma, but upon further investigation it was turning into a single vfmadd instruction a lot less than I thought, in many cases ending up as a function call (which itself was extra slow when emulating fma on hardware that lacked it). The better strategy seems to be just saying `a*b+c`, which on gcc & icc compilers automatically turns into vfmadd when available on the hardware, and adding a clang-specific pragma ensures this behavior for clang as well. Thanks to Alex Wells of Intel for pointing this out to me (in OSL land).
lgritz
added a commit
to lgritz/OpenImageIO
that referenced
this pull request
Feb 27, 2020
I was previously trying to use std::fma, but upon further investigation it was turning into a single vfmadd instruction a lot less than I thought, in many cases ending up as a function call (which itself was extra slow when emulating fma on hardware that lacked it). The better strategy seems to be just saying `a*b+c`, which on gcc & icc compilers automatically turns into vfmadd when available on the hardware, and adding a clang-specific pragma ensures this behavior for clang as well. Thanks to Alex Wells of Intel for pointing this out to me (in OSL land).
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.
I was previously trying to use std::fma, but upon further investigation
it was turning into a single vfmadd instruction a lot less than I thought,
in many cases ending up as a function call (which itself was extra slow
when emulating fma on hardware that lacked it).
The better strategy seems to be just saying
a*b+c, which on gcc &icc compilers automatically turns into vfmadd when available on the
hardware, and adding a clang-specific pragma ensures this behavior for
clang as well.
Thanks to Alex Wells of Intel for pointing this out to me (in OSL land).