Google developers have released a new compression library called Zopfli. This library, written in C, is compatible with zlib, yet provide a better compression, more exactly 3 to 8% according to Google.
This library can be used on servers for better compression in order to save bandwidth, as well as delivering web pages faster. Since it’s fully compatible with zlib, the web browsers do not need to be changed. The only drawback is that it’s several magnitude slower than zlib, so it’s better used for static content that is compressed once, and sent over the Internet many times, and it may not be a good choice for dynamic content.
The source code is available at https://code.google.com/p/zopfli/, so let’s try it.
Get the code and build zopfli:
|
1 2 3 |
git clone https://code.google.com/p/zopfli/ cd zopfli make |
Different levels of compression are available:
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 |
./zopfli -h Usage: zopfli [OPTION]... FILE -h gives this help -c write the result on standard output, instead of disk filename + '.gz' -v verbose mode --gzip output to gzip format (default) --deflate output to deflate format instead of gzip --zlib output to zlib format instead of gzip --i5 less compression, but faster --i10 less compression, but faster --i15 default compression, 15 iterations --i25 more compression, but slower --i50 more compression, but slower --i100 more compression, but slower --i250 more compression, but slower --i500 more compression, but slower --i1000 more compression, but slower |
For testing purpose, I’ve just saved this blog as one html file (test.html – 67275 bytes) with different options to see the compression improvements and time it takes with zopfli.
Command line used for the test:
|
1 |
time ./zopfli --i15 test.html && ls-l test.html.gz |
Here are the results (The first line is the compression with gzip (gzip -9 test.html):
| Option | File Size (Bytes) | Improvement | Time (s) |
| Gzip -9 | 15337 | 0.011 | |
| i5 | 14942 | 2.58% | 0.186 |
| i10 | 14942 | 2.58% | 0.287 |
| i15 | 14942 | 2.58% | 0.394 |
| i25 | 14939 | 2.60% | 0.619 |
| i50 | 14937 | 2.61% | 1.131 |
| i100 | 14937 | 2.61% | 2.199 |
| i250 | 14934 | 2.63% | 5.335 |
| i500 | 14934 | 2.63% | 11.679 |
| i1000 | 14934 | 2.63% | 21.412 |
If you run zopfli without options it will output to gzip format with 15 iterations (i15). With this particular html file, zopfli can achieve around 2.6% better compression than gzip -9, but it’s way slower on my machine, with the fastest option being about 17x slower.
Running gzip -d test.html.gz on a file compressed with zopfli, just works as advertised.
Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress
But how does it compare to other free compression bzip2, or 7zip-lzma?
They’d be better off just supporting a more modern compressor in the client, they could easily get mozilla on-board which would stick up the majority of the web.
The improvement just from the major segment of the browser market would probably more in total than just this modest improvement.
e.g. lzma gets nearly 9% on the front page, at about 450ms (unusually high sys time though)
I did not see it yesterday, but Google published their benchmark results: https://zopfli.googlecode.com/files/Data_compression_using_Zopfli.pdf