https://github.com/liquidev/pibench2
A dead-simple, multi-threaded performance benchmark.
https://github.com/liquidev/pibench2
Last synced: about 2 months ago
JSON representation
A dead-simple, multi-threaded performance benchmark.
- Host: GitHub
- URL: https://github.com/liquidev/pibench2
- Owner: liquidev
- License: mit
- Created: 2019-03-22T17:15:23.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2019-03-22T17:45:33.000Z (about 6 years ago)
- Last Synced: 2025-01-20T14:39:45.914Z (3 months ago)
- Language: Nim
- Size: 63.5 KB
- Stars: 9
- Watchers: 1
- Forks: 1
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# pibench2
A dead-simple, multi-threaded performance benchmark.
Approximates π using the (pretty inefficient) Leibniz infinite series.
## Installing
Prerequisites:
- [Nim](http://nim-lang.org/)
To install pibench2, use the following command:
```sh
$ nimble install https://github.com/liquid600pgm/pibench2
```
## Usage
pibench2 can be launched without any parameters:
```sh
$ pibench2
```
This will run the benchmark, computing 10 digits of π, using as many processor
threads as possible.
pibench2 can be tweaked using the following command line parameters:
```
--d:x --digits:x
Makes pibench2 compute x digits of π instead of the default value (10).
--t:x --threads:x
Makes pibench2 run on x threads instead of all of the CPU's threads.
Do note that this may, and probably will decrease performance if the amount
specified is greater than the CPU's thread count.
```
## Contributing
pibench2 started as a small, and very simple benchmark, but any contributions
willing to make the benchmark better, or more efficient, are welcome.
Please keep in mind the following principles while contributing:
- adhere to the [Nim Style Guide](https://nim-lang.org/docs/nep1.html),
- don't make unneccessary changes that make the codebase more convoluted,
- try not to split pibench2 into multiple modules, at least not yet,
- don't make the CLI ugly, it's supposed to be nice and modern, even if it
makes the benchmark less efficient (it doesn't).
Right now the goals of pibench2 are:
- improve the (currently) very messy codebase,
- make it more modular,
- add more (efficient) π computing algorithms.