Change where nd2 data gets copied. #853
Merged
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
We are using the dask interface with nd2 to avoid loading all of a file into memory. When we read anything other then an entire frame, the default is to copy the entire frame in memory and THEN apply whatever index and step is being used. The nd2 library docs caution that if copy isn't done, there can be a segfault. Instead, use our own lock, get the frame data, subset it with our index and step and THEN copy it; since this is in a singular lock, it won't have any greater risk that the nd2 library. Since the copy only applies to the actual data we are using, much less memory is copied. In a simple example with a file with large tiles, this saves close to 2 seconds per tile.
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Outdated suggestions cannot be applied.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.
We are using the dask interface with nd2 to avoid loading all of a file into memory. When we read anything other then an entire frame, the default is to copy the entire frame in memory and THEN apply whatever index and step is being used. The nd2 library docs caution that if copy isn't done, there can be a segfault. Instead, use our own lock, get the frame data, subset it with our index and step and THEN copy it; since this is in a singular lock, it won't have any greater risk that the nd2 library. Since the copy only applies to the actual data we are using, much less memory is copied.
In a simple example with a file with large tiles, this saves close to 2 seconds per tile.