Activate the FFI implementation of SVD on GPU. #24211
Merged
+1,072
−344
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
![copybara-service[bot]](https://avatars.githubusercontent.com/in/44061?s=60&v=4){kind=small}
copybara-service
bot
force-pushed
the
test_679137575
branch
from
4114c74
to
33d5675
Compare
copybara-service
bot
force-pushed
the
test_679137575
branch
3 times, most recently
from
e61cfa0
to
32991ce
Compare
Alongside activating this new implementation, this change adds a new `algorithm` parameter to `jax.lax.svd`. Previously the choice of algorithm was made based on heuristics in the lowering rule, but it probably also makes sense to expose an option for users to specify the algorithm explicitly because our heuristics are not very carefully optimized. This change updates the implementation of SVD in `lax` to use the FFI version which was added to jaxlib in #23794. This comes with a few benefits: 1. When running on a CUDA platform, the 64-bit API will be used for the algorithm based on QR decomposition. (Note that it looks like the 64-bit API isn't available on ROCm.) This addresses part of the feature request in #23413, although there's still work to do to port the rest of the GPU calls to the 64-bit API. 2. This implementation supports shape polymorphism in all dimensions with some caveats. By default, we do use some heuristics to based on the matrix sizes to select the algorithm that is used, and the three different algorithms (QR, Jacobi, and batched Jacobi) have sufficiently different behavior (QR returns V^H, whereas Jacobi returns V; batched Jacobi doesn't support `full_matrices=False`) that I couldn't work out a simple way to push this logic into the kernel. If the symbolic constraints are not sufficient to concretely determine the heuristics, we always use the QR algorithm. But, I've also exposed the algorithm selection in the user API, so it's possible to bypass the heuristics and get consistent behavior alongside shape polymorphism if needed. Besides these core changes, I removed the forward compatibility checks from the CPU lowering, since we're well outside of the forward compatibility window now. PiperOrigin-RevId: 687106965
copybara-service
bot
force-pushed
the
test_679137575
branch
from
32991ce
to
8361eb5
Compare
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.
Activate the FFI implementation of SVD on GPU.
Alongside activating this new implementation, this change adds a new
algorithmparameter tojax.lax.svd. Previously the choice of algorithm was made based on heuristics in the lowering rule, but it probably also makes sense to expose an option for users to specify the algorithm explicitly because our heuristics are not very carefully optimized.This change updates the implementation of SVD in
laxto use the FFI version which was added to jaxlib in #23794. This comes with a few benefits:When running on a CUDA platform, the 64-bit API will be used for the algorithm based on QR decomposition. (Note that it looks like the 64-bit API isn't available on ROCm.) This addresses part of the feature request in CuSolver: Switch to 64 bit api to allow for eigh on matrices > than 26732x26732 #23413, although there's still work to do to port the rest of the GPU calls to the 64-bit API.
This implementation supports shape polymorphism in all dimensions with some caveats. By default, we do use some heuristics to based on the matrix sizes to select the algorithm that is used, and the three different algorithms (QR, Jacobi, and batched Jacobi) have sufficiently different behavior (QR returns V^H, whereas Jacobi returns V; batched Jacobi doesn't support
full_matrices=False) that I couldn't work out a simple way to push this logic into the kernel. If the symbolic constraints are not sufficient to concretely determine the heuristics, we always use the QR algorithm. But, I've also exposed the algorithm selection in the user API, so it's possible to bypass the heuristics and get consistent behavior alongside shape polymorphism if needed.Besides these core changes, I removed the forward compatibility checks from the CPU lowering, since we're well outside of the forward compatibility window now.