[NDTensors] Circumvent scalar indexing to improve GPU performance #1216
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
kmp5VT
changed the title
…rs.jl into kmp5/enhancements/more_gpu
mtfishman
reviewed
Oct 23, 2023
mtfishman
reviewed
Oct 23, 2023
Comment on lines
+2
to
+4
| ## TODO here it looks at the elements of S so convert to CPU when on GPU | ||
| ## Could write this as a GPU impl which just converts S to array. S | ||
| ## is not used again so we don't need to convert back to GPU. |
There was a problem hiding this comment.
Same suggestion here as above for truncate!!, I think we should use leaf_parenttype dispatch.
mtfishman
reviewed
Oct 23, 2023
| D[n] = R[n, n] | ||
| end | ||
| n = size(R)[1] | ||
| D = diag(R, (n - Nd)) |
There was a problem hiding this comment.
I'm not sure this is doing what you think it is doing, diag(M, k) selects the kth off-diagonal of the matrix.
I went with:
D[1:Nd] = diag(R)[1:Nd]
mtfishman
changed the title
|
A lot of this PR is superseded by #1215. |
|
@mtfishman yes I'll close this PR and open a new one if I find something awry. Thanks! |
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Outdated suggestions cannot be applied.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.
Description
As a continued effort to make the ITensors efficient on GPU in this branch I am identifying and fixing areas that were missed in previous PRs. First I am looking at the profiling of GPU based SVD and will also address issues in 1193
TODO