Conversation
|
Your PR no longer requires formatting changes. Thank you for your contribution! |
0669d08 to
958dd36
Compare
fad3f4a to
daeab4d
Compare
|
I think this is ready for review if CI passes, it's not testing the Diagonal stuff for now waiting on @lkdvos 's PR, but it will radically simplify adding tests for that and CUDA and Enzyme |
lkdvos
left a comment
There was a problem hiding this comment.
I left a bunch of smaller comments through the code, and only quickly went through the actual test suites. Am I correct that these haven't changed much?
I think in general it would be nice to try and link some of the GPU issues we are facing with actual issue numbers and resolve them through dispatch alternatives, rather than having to alter the code for regular arrays as well.
In particular, isa(A, Array) seems a bit to strict since we also support views etc, which would then also take the slow path.
|
Now that the enzyme stuff is merged, I'll work on incorporating that into this PR and addressing the remaining comments |
|
OK, the Enzyme tests are now part of the testsuite. I turned off the |
|
OK, turned off GPU tests for Enzyme also because of EnzymeAD/Enzyme.jl#2676. I'd rather get this in and incrementally add various things as we go. |
lkdvos
left a comment
There was a problem hiding this comment.
Up to the last minor comment about the ad_utils file, I agree this is ready to go. Thanks for the large refactor!
|
It looks like the enzyme tests still need some work here |
|
Something bad is going on with |
This lets us reuse a lot of the setup infrastructure for ChainRules, Mooncake, and (soon) Enzyme. Also starts testing the AD rules on GPU.