use libdeflate rather than zlib #769
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
The zlib implementation that ships with most distributions is fairly slow, even when you allow for the zlib algorithm itself being quite slow (vs lz4 and zstd). There are faster implementations. `libdeflate` [1] is one such example. According to benchmarks [2], it's maybe 2-3x faster than zlib. This PR updates helper.cpp's compression routines to use libdeflate. It saves ~2-3% of total execution time for me: Creating a GB mbtiles (zlib): ``` real 2m1.706s user 28m24.186s sys 0m41.886s ``` Creating a GB mbtiles (libdeflate): ``` real 1m58.450s user 27m32.579s sys 0m51.848s ``` Snapshotting an external dependency is sorta distasteful - I worry about how much is too much from a maintenance POV. To mitigate that, the snapshot is a direct copy of the upstream folders, so it should be easy to update in the future if needed. This also lets us drop the zlib1g-devel and boost-iostreams dependencies, which is nice. [1]: https://github.com/ebiggers/libdeflate [2]: zlib-ng/zlib-ng#1486
Oops: if you ever passed a non-default compression level, we'd try to create a new compressor. But because we were using the default constructors, we'd happily create a temporary, copy its pointer over to the `compressor` field, then destruct the temporary, freeing the underlying native resource, and the world would end. Instead, delete the default constructors and require an explicit `setLevel` call. Also update tests to verify that the different levels work.
|
Thanks - useful improvement! Tested on Ubuntu, Intel Mac and ARM Mac and happily compiles and runs on all three. |
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.
The zlib implementation that ships with most distributions is fairly slow, even when you allow for the zlib algorithm itself being quite slow (vs lz4 and zstd).
There are faster implementations.
libdeflate1 is one such example. According to benchmarks 2, it's maybe 2-3x faster than zlib.This PR updates helper.cpp's compression routines to use libdeflate.
It saves ~2-3% of total execution time for me:
Creating a GB mbtiles (zlib):
Creating a GB mbtiles (libdeflate):
Snapshotting an external dependency is sorta distasteful - I worry about how much is too much from a maintenance POV. To mitigate that, the snapshot is a direct copy of the upstream folders, so it should be easy to update in the future if needed.
This also lets us drop the zlib1g-devel and boost-iostreams dependencies, which is nice.