2016-10-31 22:48:58 +00:00
__Zstandard__ , or `zstd` as short version, is a fast lossless compression algorithm,
2016-07-09 20:56:12 +00:00
targeting real-time compression scenarios at zlib-level and better compression ratios.
2015-01-24 00:58:16 +00:00
2016-10-31 22:48:58 +00:00
It is provided as an open-source BSD-licensed **C** library,
2017-02-02 21:50:07 +00:00
and a command line utility producing and decoding `.zst` and `.gz` files.
2016-07-12 07:54:42 +00:00
For other programming languages,
you can consult a list of known ports on [Zstandard homepage ](http://www.zstd.net/#other-languages ).
2015-01-24 00:58:16 +00:00
2017-05-05 22:41:23 +00:00
| dev branch status |
2017-05-05 22:48:42 +00:00
|-------------------|
2017-05-05 22:55:03 +00:00
| [![Build Status][travisDevBadge]][travisLink] [![Build status][AppveyorDevBadge]][AppveyorLink] [![Build status][CircleDevBadge]][CircleLink]
2017-05-05 22:48:42 +00:00
[travisDevBadge]: https://travis-ci.org/facebook/zstd.svg?branch=dev "Continuous Integration test suite"
[travisLink]: https://travis-ci.org/facebook/zstd
[AppveyorDevBadge]: https://ci.appveyor.com/api/projects/status/xt38wbdxjk5mrbem/branch/dev?svg=true "Windows test suite"
[AppveyorLink]: https://ci.appveyor.com/project/YannCollet/zstd-p0yf0
2017-05-05 22:55:03 +00:00
[CircleDevBadge]: https://circleci.com/gh/facebook/zstd/tree/dev.svg?style=shield "Short test suite"
[CircleLink]: https://circleci.com/gh/facebook/zstd
2017-05-05 22:48:42 +00:00
2015-01-24 00:58:16 +00:00
2017-03-08 00:24:54 +00:00
As a reference, several fast compression algorithms were tested and compared
2017-03-17 22:33:09 +00:00
on a server running Linux Debian (`Linux version 4.8.0-1-amd64`),
2017-03-08 00:24:54 +00:00
with a Core i7-6700K CPU @ 4.0GHz,
2017-03-17 21:25:57 +00:00
using [lzbench], an open-source in-memory benchmark by @inikep
2017-03-08 00:24:54 +00:00
compiled with GCC 6.3.0,
on the [Silesia compression corpus].
2016-02-05 14:24:57 +00:00
2017-03-17 21:25:57 +00:00
[lzbench]: https://github.com/inikep/lzbench
2016-02-05 14:24:57 +00:00
[Silesia compression corpus]: http://sun.aei.polsl.pl/~sdeor/index.php?page=silesia
2017-03-07 01:20:44 +00:00
| Compressor name | Ratio | Compression| Decompress.|
| --------------- | ------| -----------| ---------- |
| **zstd 1.1.3 -1** | 2.877 | 430 MB/s | 1110 MB/s |
| zlib 1.2.8 -1 | 2.743 | 110 MB/s | 400 MB/s |
| brotli 0.5.2 -0 | 2.708 | 400 MB/s | 430 MB/s |
| quicklz 1.5.0 -1 | 2.238 | 550 MB/s | 710 MB/s |
| lzo1x 2.09 -1 | 2.108 | 650 MB/s | 830 MB/s |
| lz4 1.7.5 | 2.101 | 720 MB/s | 3600 MB/s |
| snappy 1.1.3 | 2.091 | 500 MB/s | 1650 MB/s |
| lzf 3.6 -1 | 2.077 | 400 MB/s | 860 MB/s |
2015-08-07 19:15:27 +00:00
2015-11-02 01:36:10 +00:00
[zlib]:http://www.zlib.net/
2016-02-05 14:24:57 +00:00
[LZ4]: http://www.lz4.org/
2015-01-24 00:58:16 +00:00
2016-06-16 12:08:48 +00:00
Zstd can also offer stronger compression ratios at the cost of compression speed.
2016-09-05 23:59:26 +00:00
Speed vs Compression trade-off is configurable by small increments. Decompression speed is preserved and remains roughly the same at all settings, a property shared by most LZ compression algorithms, such as [zlib] or lzma.
2015-01-24 00:58:16 +00:00
2017-03-17 21:25:57 +00:00
The following tests were run
2017-03-17 22:33:09 +00:00
on a server running Linux Debian (`Linux version 4.8.0-1-amd64`)
2017-03-17 21:25:57 +00:00
with a Core i7-6700K CPU @ 4.0GHz,
using [lzbench], an open-source in-memory benchmark by @inikep
compiled with GCC 6.3.0,
on the [Silesia compression corpus].
2015-11-02 11:17:39 +00:00
Compression Speed vs Ratio | Decompression Speed
2015-11-02 01:44:43 +00:00
---------------------------|--------------------
2016-10-25 08:05:20 +00:00
 | 
2016-02-17 17:03:17 +00:00
2016-09-05 23:59:26 +00:00
Several algorithms can produce higher compression ratios, but at slower speeds, falling outside of the graph.
2016-10-25 08:05:20 +00:00
For a larger picture including very slow modes, [click on this link ](doc/images/DCspeed5.png ) .
2015-11-02 01:44:43 +00:00
2015-01-24 00:58:16 +00:00
2016-02-05 14:24:57 +00:00
### The case for Small Data compression
2015-01-24 09:31:47 +00:00
2017-02-02 21:50:07 +00:00
Previous charts provide results applicable to typical file and stream scenarios (several MB). Small data comes with different perspectives.
2016-02-05 14:24:57 +00:00
2017-02-02 21:50:07 +00:00
The smaller the amount of data to compress, the more difficult it is to compress. This problem is common to all compression algorithms, and reason is, compression algorithms learn from past data how to compress future data. But at the beginning of a new data set, there is no "past" to build upon.
2016-02-05 14:24:57 +00:00
2017-02-02 21:50:07 +00:00
To solve this situation, Zstd offers a __training mode__ , which can be used to tune the algorithm for a selected type of data.
Training Zstandard is achieved by provide it with a few samples (one file per sample). The result of this training is stored in a file called "dictionary", which must be loaded before compression and decompression.
Using this dictionary, the compression ratio achievable on small data improves dramatically.
2016-02-05 14:24:57 +00:00
2017-02-06 18:01:06 +00:00
The following example uses the `github-users` [sample set ](https://github.com/facebook/zstd/releases/tag/v1.1.3 ), created from [github public API ](https://developer.github.com/v3/users/#get-all-users ).
2017-02-02 21:50:07 +00:00
It consists of roughly 10K records weighting about 1KB each.
2016-02-05 14:24:57 +00:00
2017-02-02 21:50:07 +00:00
Compression Ratio | Compression Speed | Decompression Speed
------------------|-------------------|--------------------
 |  | 
2016-02-05 14:24:57 +00:00
2017-02-02 21:50:07 +00:00
These compression gains are achieved while simultaneously providing _faster_ compression and decompression speeds.
Training works if there is some correlation in a family of small data samples. The more data-specific a dictionary is, the more efficient it is (there is no _universal dictionary_ ).
Hence, deploying one dictionary per type of data will provide the greatest benefits.
Dictionary gains are mostly effective in the first few KB. Then, the compression algorithm will gradually use previously decoded content to better compress the rest of the file.
2016-02-16 15:06:53 +00:00
#### Dictionary compression How To :
1) Create the dictionary
`zstd --train FullPathToTrainingSet/* -o dictionaryName`
2016-04-12 21:58:52 +00:00
2) Compress with dictionary
2016-02-16 15:06:53 +00:00
2017-02-02 21:50:07 +00:00
`zstd -D dictionaryName FILE`
2016-02-16 15:06:53 +00:00
3) Decompress with dictionary
2017-02-02 21:50:07 +00:00
`zstd -D dictionaryName --decompress FILE.zst`
2016-02-16 15:06:53 +00:00
2016-09-19 12:58:14 +00:00
### Build
Once you have the repository cloned, there are multiple ways provided to build Zstandard.
#### Makefile
2016-09-21 09:24:22 +00:00
If your system is compatible with a standard `make` (or `gmake` ) binary generator,
you can simply run it at the root directory.
It will generate `zstd` within root directory.
2016-09-19 12:58:14 +00:00
Other available options include :
- `make install` : create and install zstd binary, library and man page
- `make test` : create and run `zstd` and test tools on local platform
#### cmake
A `cmake` project generator is provided within `build/cmake` .
It can generate Makefiles or other build scripts
to create `zstd` binary, and `libzstd` dynamic and static libraries.
2017-02-17 10:29:23 +00:00
#### Meson
2017-02-18 21:30:57 +00:00
A Meson project is provided within `contrib/meson` .
2017-02-17 10:29:23 +00:00
2017-02-27 10:58:22 +00:00
#### Visual Studio (Windows)
2016-09-19 12:58:14 +00:00
Going into `build` directory, you will find additional possibilities :
- Projects for Visual Studio 2005, 2008 and 2010
+ VS2010 project is compatible with VS2012, VS2013 and VS2015
- Automated build scripts for Visual compiler by @KrzysFR , in `build/VS_scripts` ,
which will build `zstd` cli and `libzstd` library without any need to open Visual Studio solution.
2016-02-05 14:24:57 +00:00
### Status
2016-09-05 23:59:26 +00:00
Zstandard is currently deployed within Facebook. It is used daily to compress and decompress very large amounts of data in multiple formats and use cases.
2016-08-30 17:04:33 +00:00
Zstandard is considered safe for production environments.
### License
Zstandard is [BSD-licensed ](LICENSE ). We also provide an [additional patent grant ](PATENTS ).
2015-01-24 00:58:16 +00:00
2016-08-30 17:04:33 +00:00
### Contributing
2016-02-05 14:24:57 +00:00
2016-08-30 17:04:33 +00:00
The "dev" branch is the one where all contributions will be merged before reaching "master".
If you plan to propose a patch, please commit into the "dev" branch or its own feature branch.
Direct commit to "master" are not permitted.
For more information, please read [CONTRIBUTING ](CONTRIBUTING.md ).
2016-02-05 14:24:57 +00:00
2016-04-12 21:58:52 +00:00
### Miscellaneous
2016-02-05 14:24:57 +00:00
Zstd entropy stage is provided by [Huff0 and FSE, from Finite State Entropy library ](https://github.com/Cyan4973/FiniteStateEntropy ).
