|Project Name||Stars||Downloads||Repos Using This||Packages Using This||Most Recent Commit||Total Releases||Latest Release||Open Issues||License||Language|
|Mango||913||14||2||2 years ago||37||July 14, 2020||24||apache-2.0||Java|
|Distributed ORM Framework for Java|
|Hikaricp Benchmark||137||2 years ago||18||Java|
|JHM benchmarks for JDBC Connection Pools|
|Tpcc||25||7 years ago||1||Java|
|Java implementation of TPC-C benchmark|
|Activiti Benchmark||21||8 years ago||Java|
|Shardingsphere Benchmark||17||2 years ago||1||Java|
|Distributed database middleware|
|Mango Benchmark||15||7 years ago||Java|
|JHM benchmarks for ORM Frameworks|
|Sqlaction||14||a year ago||8||May 25, 2019||2||apache-2.0||Java|
|Database persistence layer tool based auto-gen JDBC code|
|Dbms Benchmarker||10||1||15 days ago||16||June 29, 2022||20||agpl-3.0||HTML|
|DBMS-Benchmarker is a Python-based application-level blackbox benchmark tool for Database Management Systems (DBMS). It connects to a given list of DBMS (via JDBC) and runs a given list of parametrized and randomized (SQL) benchmark queries. Evaluations are available via a Python interface and on an interactive multi-dimensional dashboard.|
|Benchmarksql||7||3 years ago||2||Java|
|BenchmarkSQL PostgreSQL TPC-C fair-use implementation (unmaintained, new maintainers welcome)|
|Sql Benchmark||4||4 years ago||apache-2.0||Ada|
|Tool to make SQL benchmark on different drivers, languages and databases|
ConnectionCycle measures cycles of
DataSource.getConnection()/Connection.close(). StatementCycle measures cycles of
This set of microbenchmaks was developed to refine the HikariCP JDBC connection pool implementation, but it actually runs the same benchmarks across multiple pools.
We have come to understand that benchmarking on the JVM, which employs Dead Code Elimination (DCE), lock-coalescing, inlining, loop-unrolling, on-stack replacement (OSR) and a myriad of other tricks, renders most attempts at benchmarking completely invalid -- including our own original benchmarks. Read all the things that [even smart] people get wrong about benchmarking on the JVM.
The Oracle JVM performance team, primarily Aleksey Shipilёv, developed a microbenchmarking framework called JMH. It provides the infrastructure (if used properly) for accurate comparative measurement of JVM-based execution. If you are interested in microbenchmarking at all, or just curious about all the wonderful things the JVM does, I highly recommend reading this slideshare.
git clone https://github.com/brettwooldridge/HikariCP-benchmark.git
mvn clean package
benchmark.sh script is a wrapper around JMH execution. A full run of the benchmark will take about 45 minutes for all pools.
There are several more options you can provide to the
benchmark.sh. There are a lot actually, but these are most useful...
Specify Shorter Runs
There are two options provided by the script:
medium. quick will take about 5 minutes to run, medium will take about 20 minutes -- for all pools. It is extrememly boring to watch, and you can't do anything else on the PC where the benchmark is running without affecting the results, so have dinner, run some errands, etc.
If specified with other options,
medium must be the first option.
Specify Specific Pools
./benchmark.sh -p pool=hikari,bone
pool is a comma-separated list (hikari, dbcp2, tomcat, c3p0, vibur). Specifying a specific pool or subset of pools will shorten run times.
Specify Pool Size
./benchmark.sh -p maxPoolSize=16
Pool size is only applicable for the Connection Cycle test, attempting to run the Statement Cycle test with a pool smaller than the number of threads (8) will result in testing failures. The Connection Cycle test runs with 8 threads, so to test a contrained pool condition set maxPoolSize to a smaller number (eg. 4).
Specify which Benchmark
There are two benchmarks in the suite currently: ConnectionBench and StatementBench. By default both benchmarks are run, but if you want to run one or the other you can use a JMH option using a regex (regular experession) to do so. For example, to only run the StatementBench use:
All of the options can be combined:
./benchmark.sh medium -p pool=hikari,vibur -p maxPoolSize=4 ".*Connection.*"