https://github.com/fasterxml/jackson-benchmarks
Project that contains JMH-based micro-benchmarks to help with optimizations
https://github.com/fasterxml/jackson-benchmarks
Last synced: 6 days ago
JSON representation
Project that contains JMH-based micro-benchmarks to help with optimizations
- Host: GitHub
- URL: https://github.com/fasterxml/jackson-benchmarks
- Owner: FasterXML
- Created: 2014-05-09T05:26:21.000Z (almost 11 years ago)
- Default Branch: 2.x
- Last Pushed: 2025-04-25T02:48:09.000Z (8 days ago)
- Last Synced: 2025-04-25T03:34:19.132Z (8 days ago)
- Language: Java
- Size: 800 KB
- Stars: 16
- Watchers: 4
- Forks: 17
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Overview
This project contains simple set of performance micro-benchmarks, using excellent
[JMH](http://openjdk.java.net/projects/code-tools/jmh/) package.
# Status
Has been used since Jackson 2.4 to track performance changes: see various `results-pojo-2.x-home.txt`
files for numbers.
## Usage
Tests usually run with
run-mediaitem.sh
or directly with something like
java -Xmx256m -jar target/microbenchmarks.jar ".*StdReadVanilla.*PojoMediaItem.*" -wi 4 -w 1 -i 5 -r 1 -f 3 -t 1
(see `results-pojo-2.13-home.txt` and others for specific invocations used for each group of numbers)
## Profiling
Can use
./profile-mediaitem.sh
to use basic JMH settings over runs.
## Test sets
### POJO: MediaItem
`MediaItem` POJO comes from: https://github.com/eishay/jvm-serializers tests suite.
Test input generated is about 300 bytes as JSON, and between 200 and 500 bytes across other formats (converted from JSON which is considered canonical).
Settings to run tests include:
* Standard: java -Xmx256m -jar target/microbenchmarks.jar ".*StdReadVan.*PojoMedia.*" -wi 4 -w 1 -i 5 -r 1 -f 9 -t 1
* Afterburner: java -Xmx256m -jar target/microbenchmarks.jar ".*StdReadAfter.*PojoMedia.*" -wi 4 -w 1 -i 5 -r 1 -f 9 -t
1
### POJO: Currency (floating-point (double))
`Currency` POJO, and data file `./json/USD.json` (about 2kB) come from [Awesome JSON Datasets](https://github.com/jdorfman/awesome-json-datasets#currency).
Settings to run tests include:
```
* Default/JSON: java -Xmx256m -jar target/microbenchmarks.jar ".*JsonStdReadVanilla.readCurrencyPojoDefault.*" -wi 4 -w 1 -i 5 -r 1 -f 3 -t 1
* FastFP/JSON: java -Xmx256m -jar target/microbenchmarks.jar ".*JsonStdReadVanilla.readCurrencyPojoFast.*" -wi 4 -w 1 -i 5 -r 1 -f 3 -t 1
* All: java -Xmx256m -jar target/microbenchmarks.jar ".*StdReadVanilla.readCurrencyPojo.*" -wi 4 -w 1 -i 5 -r 1 -f 3 -t 1
```
### POJO: Currency (BigDecimal)
Same as above, but instead of reading as `double` values are read as `BigDecimal`. Run using
```
* Default/JSON: java -Xmx256m -jar target/microbenchmarks.jar ".*JsonStdReadVanilla.readCurrencyBigDefault.*" -wi 4 -w 1 -i 5 -r 1 -f 3 -t 1
* FastFP/JSON: java -Xmx256m -jar target/microbenchmarks.jar ".*JsonStdReadVanilla.readCurrencyBigFast.*" -wi 4 -w 1 -i 5 -r 1 -f 3 -t 1
* All: java -Xmx256m -jar target/microbenchmarks.jar ".*StdReadVanilla.readCurrencyBig.*" -wi 4 -w 1 -i 5 -r 1 -f 3 -t 1
```