concurrency-benchmarks: Benchmarks to compare concurrency APIs
Benchmarks to compare the pure concurrency overhead of various flavors of
concurrent streamly streams and the async package.
Use cabal new-bench or stack bench to run the benchmarks. To generate
charts, run the benchmarks with --csv-raw=results.csv option and then run
makecharts results.csv. Charts are generated in the charts directory.
[Skip to Readme]
Downloads
- concurrency-benchmarks-0.1.0.tar.gz [browse] (Cabal source package)
- Package description (as included in the package)
Maintainer's Corner
For package maintainers and hackage trustees
Candidates
- No Candidates
| Versions [RSS] | 0.1.0, 0.1.1 |
|---|---|
| Change log | Changelog.md |
| Dependencies | base (>=4 && <5), bench-graph (>=0.1 && <0.2), bytestring (>=0.9 && <0.11), Chart (>=1.6 && <2), Chart-diagrams (>=1.6 && <2), csv (>=0.1 && <0.2), directory (>=1.2 && <1.4), getopt-generics (>=0.11 && <0.14), split (>=0.2 && <0.3), text (>=1.1.1 && <1.3), transformers (>=0.4 && <0.6), typed-process (>=0.1.0.0 && <0.3) [details] |
| License | MIT |
| Copyright | Copyright (c) 2018 Harendra Kumar |
| Author | Harendra Kumar |
| Maintainer | Harendra Kumar |
| Category | Benchmark |
| Home page | http://github.com/composewell/concurrency-benchmarks |
| Bug tracker | http://github.com/composewell/concurrency-benchmarks/issues |
| Source repo | head: git clone git://github.com/composewell/concurrency-benchmarks.git |
| Uploaded | by harendra at 2018-09-06T17:09:36Z |
| Distributions | |
| Reverse Dependencies | 1 direct, 0 indirect [details] |
| Executables | makecharts |
| Downloads | 1250 total (12 in the last 30 days) |
| Rating | (no votes yet) [estimated by Bayesian average] |
| Your Rating | |
| Status | Docs not available [build log] Last success reported on 2018-09-06 [all 3 reports] |
Readme for concurrency-benchmarks-0.1.0
[back to package description]concurrency-benchmarks
Benchmarks to compare the pure concurrency overhead of various flavors of concurrent streamly streams and the async package.
Use cabal new-bench or stack bench to run the benchmarks. To generate
charts, run the benchmarks with --csv-raw=results.csv option and then run
makecharts results.csv. Charts are generated in the charts directory.
Methodology
A total of 100,000 tasks are run for each concurrency mechanism being compared. Each task is a noop i.e. it does nothing, just returns. Therefore the benchmark measures the pure overhead of concurrency for the tiniest possible tasks.
Streamly's async and ahead style streams may automatically run the tasks in
less number of threads than the actual number of tasks i.e. they may run
multiple tasks per thread, if the tasks do not block and have a very low
latency. Therefore these streams are much faster on this benchmark. However,
the streamly parallel style stream guarantees that each task runs in a
separate thread however small it is. For the async package,
mapConcurrently_ is used, which runs the tasks but does not collect the
results.
Results
These charts compare the streamly master
branch
and async-2.2.1 on a MacBook Pro with a 2.2 GHz Intel Core i7 processor.
This chart shows the time taken for the benchmark completion.
This chart shows the maximum rss (resident set size), in other words peak memory consumed during the benchmark run.
Feedback
Feedback is welcome. Please raise an issue, send a PR or send an email to the author.