Batched linear system of equations solver (torch.bgesv) #4502
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
Collaborator
Contributor
|
cc @fritzo |
Collaborator
apaszke
reviewed
Jan 5, 2018
Contributor
apaszke
left a comment
apaszke
left a comment
There was a problem hiding this comment.
That looks great, thanks for wrapping this up so quickly! Only a few minor comments.
| } | ||
|
|
||
| magma_queue_t magma_queue; | ||
| magma_queue_create_from_cuda( |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
| THError("MAGMA bgesv (gesv_batched) : For batch number %lld: U(%d,%d) is zero, singular U.", | ||
| (long long)batch_count, info, info); | ||
| } | ||
| } |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
| THTensor_(resizeNd)(self, 3, size, stride); | ||
| THTensor_(copy)(self, src); | ||
| return self; | ||
| } |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
| x_exp, LU_exp = torch.gesv(b.squeeze(0), A.squeeze(0)) | ||
| x, LU = torch.bgesv(b, A) | ||
| self.assertEqual(x, x_exp.unsqueeze(0)) | ||
| self.assertEqual(LU, LU_exp.unsqueeze(0)) |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
- Refactored THCTensor_(newBatchedColumnMajor) and THTensor_(cloneBatchedColumnMajor) - In THCTensor_(bgesv), move error checking to after freeing a lot of things - Added a test to check bgesv against gesv (in a loop) for a batch of size 4
Member
|
A few high-level comments: The equivalent NumPy function is np.linalg.solve
You don't have to block the PR on these suggestions, but if it's not too much extra work see if you can merge gesv and bgesv so that we don't unnecessarily expand the public API. |
Contributor
Author
|
Closing this in favor of #4612 |
zou3519
added a commit
to zou3519/pytorch
that referenced
this pull request
May 7, 2018
Fixes pytorch#3164 Picks up from pytorch#4502 I moved `gesv` to ATen. Adds bindings for MAGMA's `gesv_batched` function for CUDA. For CPU, runs `THLapack(gesv)` in a for loop. The new function supports arbitrary batch dimensions (and broadcasting of those dimensions). For example, the 4-d tensor `A x B x M x M` should be treated as having batch-size `(A x B)`. The overhead of creating the magma_queue_t is: ~350000 microseconds the first time it's called and ~6 microseconds every time after that.
soumith
pushed a commit
that referenced
this pull request
May 8, 2018
* Add batched linear solver to torch.gesv() Fixes #3164 Picks up from #4502 I moved `gesv` to ATen. Adds bindings for MAGMA's `gesv_batched` function for CUDA. For CPU, runs `THLapack(gesv)` in a for loop. The new function supports arbitrary batch dimensions (and broadcasting of those dimensions). For example, the 4-d tensor `A x B x M x M` should be treated as having batch-size `(A x B)`. The overhead of creating the magma_queue_t is: ~350000 microseconds the first time it's called and ~6 microseconds every time after that. * Tests and docs * Address comments * Address comments * Rebase * Address comments * Fix rebase * Addressed comments * Address comments * Address comments * Addressed comments
weiyangfb
pushed a commit
to weiyangfb/pytorch
that referenced
this pull request
Jun 11, 2018
* Add batched linear solver to torch.gesv() Fixes pytorch#3164 Picks up from pytorch#4502 I moved `gesv` to ATen. Adds bindings for MAGMA's `gesv_batched` function for CUDA. For CPU, runs `THLapack(gesv)` in a for loop. The new function supports arbitrary batch dimensions (and broadcasting of those dimensions). For example, the 4-d tensor `A x B x M x M` should be treated as having batch-size `(A x B)`. The overhead of creating the magma_queue_t is: ~350000 microseconds the first time it's called and ~6 microseconds every time after that. * Tests and docs * Address comments * Address comments * Rebase * Address comments * Fix rebase * Addressed comments * Address comments * Address comments * Addressed comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 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.
Implements torch.bgesv, a batched linear system of equations solver.
Adds bindings for MAGMA's gesv_batched function for CUDA.
For CPU, runs THLapack(gesv) in a for loop.
I decided to not build this into
torch.gesvbut if we want to I can change that.cc @apaszke
Test Plan
New unit tests:
torch.gesvand specifying outputs