Initial upstreaming of Rust ADBench implementations #149126
Open
+2,004
−0
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
rustbot
added
S-waiting-on-review
Collaborator
|
|
This comment has been minimized.
This comment has been minimized.
ZuseZ4
force-pushed
the
autodiff-benchmark-tests
branch
from
5968770 to
78d8adf
Compare
This comment has been minimized.
This comment has been minimized.
ZuseZ4
force-pushed
the
autodiff-benchmark-tests
branch
from
78d8adf to
e29cd37
Compare
This comment has been minimized.
This comment has been minimized.
ZuseZ4
force-pushed
the
autodiff-benchmark-tests
branch
from
e29cd37 to
1fb0a0b
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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Outdated suggestions cannot be applied.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.
One of my last steps in order to prepare autodiff on nightly was to upstream those benchmarks into r-l/r,
since they can now be part of our CI runs. Those are the first real-world testcases of a relevant size which we'll have, and we ran into a lot of enzyme bugs with them in the past (and depending on the config still do).
I think they are good to have around to learn about both Enzyme regressions and improvements.
If any of them fails after an llvm or rustc update, people can simply disable the specific one.
I have a safe and unsafe (ptr based) implementation for each, since they have quite a perf difference between them (unsafe usually being slower). They also can trigger quite different bugs in Enzyme.
For the same reason I'd love to also have a version around which uses libm instead of the rust builtin function, since the libm version gets vectorized better in some of these benchmarks, which again triggered different bugs in the past.
rust-lang/rust-project-goals#109 (comment)
r? @oli-obk
Edit: I probably want to change the folder structure, so that we have all the lib.rs on a higher level, sofixed, I marked side files as ignore-aux.x.py testonly runs them, and ignores the other dependent files. I'll look into it.