Optimize TileRenderer tile image sorting #30
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
This improves overall render performance by roughly 10%. This has two main differences: - We make separate calls to std::sort for each comparator type, to avoid dynamic dispatch at runtime and allow the comparator to be inlined - We only sort pointers to the tiles instead of tiles themselves. This means substantially less data has to be moved around, as each step in the sorting algorithm only has to swap two pointers instead of repeatedly move-constructing TileImage instances, and is responsible for most of the performance improvement.
…images vector this reduces how often the vector's storage needs to be grown during rendering, which is a pretty expensive process
this saves an additional memset
Owner
|
Hi |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
This contains three main changes:
TileImages directly, we'll leave them unmodified and sort pointers to them instead. This makes the actual sorting much faster, since it only has to swap a pair of pointers after each comparison instead of a 56-byte class with a non-trivial move assignment operator.std::sort()is now invoked separately for each of the sort modes, allowing the comparator to be inlined into the sorting codetile_imagesvector when filling it. This mitigates that by reserving a generous amount of space ahead of time, and also keeps track of the global maximum number of images per tile encountered so far.These three optimizations reduce the total render time of a simple map by ~15% on my machine.
I've also replaced
boost::container::vectorwithstd::vectorjust to shave off an#include.