-
Notifications
You must be signed in to change notification settings - Fork 1.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[ntuple] Simplify and clean up compression #17847
base: master
Are you sure you want to change the base?
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Very nice, thanks!
Test Results 18 files 18 suites 4d 8h 13m 37s ⏱️ For more details on these failures, see this check. Results for commit 5423345. ♻️ This comment has been updated with latest results. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Very nice! I have one small doc-related comment but otherwise it's good to go.
InitImpl, CommitClusterGroupImpl, and CommitDatasetImpl already allocate a buffer that is large enough and it is not needed to pass via the internal buffer of RNTupleCompressor just to memcpy the compressed contents out again.
It was already unused in SealPage since commit 208061d ("fix unbuffered compression of large pages"). This saves 16 MiB of internal compression buffer per page sink.
A RNTupleDecompressor will internally allocate a buffer of 16 MiB, which is not needed for the anchor. It was also the only usage of in-place decompression with RNTupleDecompressor.
They are not used anymore in the source code.
They are not used anymore in the source code.
5423345
to
d1b534d
Compare
(just rebasing to get a green CI before merging) |
... eventually removing all member methods of RNTuple(De)Compressor; see the individual commits for more details. As a visible result, this saves 16 MiB of internal compression buffer per page sink.