go api: reduce allocations in deck decompression #19
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
This utility is pretty handy for comparing benchmarks in a more human friendly way.
This file previously did not have tests, which is probably fine, but since
I wanted to change it a bunch and have not worked with it adding tests
first seemed like a safer thing to do.
This also adds a benchmark which can show us how expensive this is
right now. The current output looks like:
```
goos: linux
goarch: amd64
pkg: github.com/MaT1g3R/slaytherelics/api
cpu: AMD Ryzen 9 5900X 12-Core Processor
│ before.txt │
│ sec/op │
DecompressDeck/Simple_deck-24 7.934µ ± 1%
DecompressDeck/Big_deck-24 227.6µ ± 1%
geomean 42.49µ
│ before.txt │
│ B/op │
DecompressDeck/Simple_deck-24 5.366Ki ± 0%
DecompressDeck/Big_deck-24 330.8Ki ± 0%
geomean 42.13Ki
│ before.txt │
│ allocs/op │
DecompressDeck/Simple_deck-24 94.00 ± 0%
DecompressDeck/Big_deck-24 1.810k ± 0%
geomean 412.5
```
Previously we would allocate quite often in order to decompress and
parse the deck strings that were given to us. This could be alleviated by
precomputing some of the regular expressions, preallocating, and using
manual conversion logic instead of relying on json.
This could obviously be taken much further, but this seemed like a good
compromise on benefit while still retaining readability.
Benchmark comparison:
```
goos: linux
goarch: amd64
pkg: github.com/MaT1g3R/slaytherelics/api
cpu: AMD Ryzen 9 5900X 12-Core Processor
│ before.txt │ after.txt │
│ sec/op │ sec/op vs base │
DecompressDeck/Simple_deck-24 7.934µ ± 1% 1.318µ ± 0% -83.39% (p=0.000 n=10)
DecompressDeck/Big_deck-24 227.58µ ± 1% 60.34µ ± 1% -73.48% (p=0.000 n=10)
geomean 42.49µ 8.916µ -79.02%
│ before.txt │ after.txt │
│ B/op │ B/op vs base │
DecompressDeck/Simple_deck-24 5494.5 ± 0% 836.0 ± 0% -84.78% (p=0.000 n=10)
DecompressDeck/Big_deck-24 330.8Ki ± 0% 219.2Ki ± 0% -33.72% (p=0.000 n=10)
geomean 42.13Ki 13.38Ki -68.24%
│ before.txt │ after.txt │
│ allocs/op │ allocs/op vs base │
DecompressDeck/Simple_deck-24 94.00 ± 0% 17.00 ± 0% -81.91% (p=0.000 n=10)
DecompressDeck/Big_deck-24 1810.0 ± 0% 165.0 ± 0% -90.88% (p=0.000 n=10)
geomean 412.5 52.96 -87.16%
```
MaT1g3R
approved these changes
Oct 10, 2024
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.
Previously we would allocate quite often in order to decompress and
parse the deck strings that were given to us. This could be alleviated by
precomputing some of the regular expressions, preallocating, and using
manual conversion logic instead of relying on json.
This could obviously be taken much further, but this seemed like a good
compromise on benefit while still retaining readability.
Benchmark comparison: