Welcome to the most extensive file compression benchmark ever. CompressionRatings.Com tests are pedantic and verifiable. All the software and test material are public.
CompressionRatings.Com is a reference for both professionals and end-users looking for the latest information about the performance of lossless data compression software and technology. Quick links
detailed benchmark results
test breakdown
comparison tool
1-min tutorial
The site provides vast amount of benchmark data. It can be used to compare relative performance of compression software and provide evidence based answers to end-user questions such as: how compressors compare, which one has "the best compression" or "the highest compression ratio", etc.
Continue to the benchmark results...
Zhuff 0.3nr submitted by the author.
Astrotite 1.5 (and 1.6 January 16) submitted by the author.
Zhuff 0.2 (and 0.2a on January 6, 0.3 on January 19) submitted by the author.
Reference files is a new feature for the site. We run tests with common test files with all compressors (including those that did not qualify for the main benchmark).
Zhuff 0.1 submitted by the author.
Bee 0.79 submitted by the author.
sr 3C submitted by the author.
LZP2 091011 submitted by the author.
M03 0.2a1 submitted by the author.
LZP2 091010, 091010b, 091010c submitted by the author.
M03 0.2a submitted by the author.
M03 0.1a submitted by the author.
Flashzip 0.99b5 submitted by the author.
Fixed an error in Application1 timings for Flashzip 0.99b4 -m1 -c7 -b8.
Changes to the corpus. Audio1 set is replaced by a 'mini' version that consists of the same samples cut in half of the original length. Now the audio has a weight of 3% and images 3% in the corpus. Text is overweighted at 9% (on the other hand text is more compressible than audio or images) considering it is unorthodox to find hundreds of megabytes of text in one place.
csc 3.1 submitted by the author.
Flashzip 0.99b4 submitted by the author.
The site has a new feature. Top 10 page shows a brief summary of the results for the best compressors by rating.
The qualification threshold is reduced again, to a 1/10 rating of the best one (previously 1/6).
bcm 0.09 submitted by the author.
Stuffit 13.0 added.
Continued to lower the qualification threshold. A rating 1/6 of the best one (previously 1/5) qualifies. Some additional compressors/configurations now qualify.
RINGS 1.6 submitted by the author.
csc 3 submitted by the author.
Flashzip 0.99b3 submitted by the author.
Packet 0.91a added.
The site has another new feature, a comparison tool. This tool helps comparing two or more compressors (or configurations). See 7-Zip vs WinRAR for example.
Since fast compressors add no significant strain to the benchmark system, the following modification is made to the qualification rules: any compressor that is faster (either compressing or decompressing) than PPMd Jr1 (default conf.) is qualified. Squeez, JAR, bwtzip, QazaR, flzp and other compressors qualify and additional configurations from XWRT, lzturbo, Flashzip, and from several others now qualify as well.
Finally the site has data visualizations. You can find graphics for each test in their summary pages. The BWT comparison has got some graphics as well.
Pareto frontier is the most requested feature, it is now implemented. Times are underlined if no faster program exist that has higher compression ratio. It's also possible to filter out other results and see only the pareto frontier. You can find the pareto frontier links for each test right above the calculator link.
Flashzip 0.99b1 submitted by the author.
KGB Archiver 2.0 beta added.
All programs have been tested again with Application1 because the database was apparently corrupted. Used this opportunity to again make changes to the test with the removal of empty directories and more small files.
BWMonstr 0.02 added.
Changes to the BWT comparison: enwik500m is replaced with enwik9, 1 new test file and 2 new transformation test files.
Changed the way qualification ratings are calculated for the program pages (i.e. not for the main result pages). Now the decoder program size is excluded from the calculations because mostly these numbers are used for program version comparison only, so this will reduce random fluctuations in the numbers.
NanoZip 0.07a added.
While testing this program it was discovered that if a program uses an additional thread while accessing disk, the disk IO time gets included with the kernel and user times. For NanoZip 0.07a some of the times for the fast modes are 80% slower what they should be. If somebody knows how to prevent this (without modifying the program) from happening please let me know.
BWTmix 1 submitted by the author.
BWTmix 0 submitted by the author.
The site has a new background. It consists of Lena, the famous compression test image and PPMd source code. If this isn't nerdy enough for you, then nothing is.
bcm 0.08 submitted by the author.
Changes to the Application1 test. Some more of the data is now decompressed (it's now 50 MB larger) and the large number of very small png files removed. Finished testing everything again because of the April 5. bios setting change. Added quick links and links to the detailed results for the rating page. Hopefully they be a help navigating the site.
Here is an article (with source code attached) for authors about implementing an archiver. Authors can use this to turn a compressor into an archiver in no time.
lzturbo 0.95 submitted by the author.
Hook 1.4 submitted by the author.
LZP2 090420 submitted by the author.
Calculator is a new feature. You can find the link for each test in the summary page. The calculator can be used to calculate optimal compressor for a protocol or used as an alternative custom user rating system. There is a save button which stores the calculator setup into a cookie, so that whenever you click the calculator link your setup will be restored.
7-Zip 9.01a added. See the program page how this version compares to the earlier ones.
PPMd sh8 submitted by the author.
LZP2 (two versions) submitted by the author.
Improvements for program pages. Now the absolute ratio and speed (kB/s) are listed for each configuration. See ZPAQ, 7-Zip, ASH pages for an example. Another improvement is for the qualification results. Gray color is used for ratings below the qualification threshold. Hover mouse over such number to see the difference.
Software added: ASH 0.7 and 0.7b1 submitted by the author.
Software added: LZP2 submitted by the author.
An additional unpacker for lzturbo -59 submitted by the author.