Summary of all single file-type lossless data compression tests


Test Best program Original Compressed Compression Bits per
  Size (bytes) Size (bytes) Ratio (%) Byte (b/B)
LOGFILEPAQ8PX2061707125719398.75 0.0998
ENGLISH TEXTWinRK 3.1.2298857833057188.94 0.8849
SORTED WORDLISTPAQ8PX406743938603290.51 0.7593
HLPPAQ8PX412141840011290.29 0.7766
MS-WORD DOC FILEPAQ8PX416819248286488.42 0.9268
BITMAPPAQ8PX414941453900387.01 1.0392
JPG/JPEGPAQ8PX84246863712424.37 6.0501
EXECUTABLEWinRK 3.1.2387078489636576.84 1.8526
DLL (EXECUTABLE)WinRK 3.1.23782416123664367.31 2.6156
PDFWinRK 3.1.24526946354919721.60 6.2721

Best programs matrix


Program/TestBMPDICTDLLDOCEXEHLPJPGLOGPDFTEXT
DURILCA 0.574190047378412816747147289452104815288294253413683633255385377
EPM r9739387491663139617073126710482385034358292433697073657343413392
LPAQ8721553411980154738673418911399124884468230053572633629026427349
NanoZip 0.09a716979443110141333669836810346744657638131123588303613181395153
PAQ8PX53900338603212928694828649091614001126371242571933556044352722
PAQAR 4.568735643567413406797059039522664297668253303151823602320333759
PIMPLE2732085559006155708074949511795105029188299404357573649152445337
PPMonstr J rev.1742068477472155860972401911724704839358293593557223635320410847
SLIM 0.23d74443846369212996947278069663944848898303763388943644735405001
WinRK 3.1.256905339370412366436882378963654155228127002716283549197330571
ZPAQ 2.05714217449923157235172828012056364841498224304014293624628416929



Best overall compression program #1


Position Program #1st pos #2nd pos #3th pos #4th pos #5th pos #6th pos #7th pos #8th pos Points
01PAQ8PX6220000088
02WinRK 3.1.24410100082
03PAQAR 4.50133200049
04NanoZip 0.09a0012123133
05DURILCA 0.50110401032
06SLIM 0.23d0002101218
07STUFFIT 140200000016
08ZPAQ 2.050000121113
09LPAQ80010011112
10PPMonstr J rev.10000030110
11RKC 1.02000100027
12EPM r9000001105
13Ocamyd 1.66test1000000204

- Scoring similar to Formula 1 2003 system:
  1st place, 10 points;  2nd place, 8 points
  3th place, 6 points ;  4th place, 5 points
  5th place, 4 points ;  6th place, 3 points
  7th place, 2 points ;  8th place, 1 point
- Only programs with at least 2 top8 scores will be listed.
- If two programs have the same amount of points the program with the most 1st places (2nd places etc) will be regarded the better program.



Best overall compression program #2


Pos # Program Total size of 10 tests Ratio
  Bytes (%)
001PAQ8PX8813124100.0
002WinRK 3.1.29163620104.0
003PAQAR 4.59628235109.2
004DURILCA 0.59828249111.5
005SLIM 0.23d9905919112.4
006NanoZip 0.09a9952506112.9
007EPM r910179845115.5
008LPAQ810280109116.6
009PPMonstr J rev.110389821117.9
010RKC 1.0210411833118.1
011ZPAQ 2.0510419972118.2
012CMM4 0.2b10514487119.3
013PIMPLE210640280120.7
014COMPRESSIA 1.0b10664630121.0
015CCM 1.30c10736835121.8
016ASH 0710821374122.8
017BIT 0.710875515123.4
018Ocamyd 1.66test110935918124.1
019ENC 0.1510939309124.1
020HIPP 0.581910956556124.3
021FreeARC 0.66611134548126.3
022UHARC 0.6b11151014126.5
023TC 5.2 dev211279556128.0
0247-Zip 9.25a11281494128.0
025STUFFIT 1411284662128.0
026BEE 0.7.911298396128.2
027RZM 0.07h11321527128.5
028SCM 0.0.1b11347720128.8
029SQUEEZ 5.6311373488129.1
030Ultra7z Opt 0.0511378813129.1
031RK 1.04.111454328130.0
032BSC 3.0.011571019131.3
033CTXf 0.75 b111587323131.5
034WinRAR 4.1b311601141131.6
035Quark 0.95r11639975132.1
036WINZIP 1411650076132.2
037PIM 2.9011666751132.4
038PPMN 1.00b1 km11733001133.1
039GRZipII 0.2.411805752134.0
040RKUC 1.0411825088134.2
041PPMVC 1.211834615134.3
042PPMY SSE (9A9)11835558134.3
043PPMd rev J11868185134.7
044777 0.04b111887994134.9
045RINGS 1.611896784135.0
046GRZIP 0.7.311915243135.2
047FlashZIP 0.99b811948292135.6
048HOOK 1.411950372135.6
049BWMonstr 0.0211967621135.8
050M1 0.3b11972113135.8
051LZPX(J) 1.2h11977759135.9
052UFA 0.04b112033036136.5
053Blizzard 0.24b12035530136.6
054BMA 1.35b12036301136.6
055ICEOWS 4.20b12094203137.2
056BSSC 0.95a12117125137.5
057RKIVE 1.9212194910138.4
058LZTurbo 0.9512204073138.5
059SZIP 1.1212212231138.6
060ACB 2.00c12215499138.6
061UHBC 1.012238088138.9
062CTW 0.112309636139.7
063SBC 0.970 rev312313311139.7
064WinHKI 1.7412352672140.2
065BALZ 1.1512364886140.3
066BCM 0.1212371596140.4
067M0312375777140.4
068DC 0.99.307b12443207141.2
069ZZIP 0.36c12472459141.5
070BioArc 1.912511636142.0
071DARK 0.5112517457142.0
072YZX 0.0412534349142.2
073BBB ver112536777142.3
074BruteCM 0.1d12574778142.7
075YBS 0.03f12578095142.7
076QC 0.05012607670143.1
077LZPM 0.1612612129143.1
078PPMZ2 0.8112628639143.3
079QLFC 6.6w12640987143.4
080Quad 1.1212647083143.5
081TarsaLZP 21Aug200712647559143.5
082MNZIP12653353143.6
083BA 1.0112654768143.6
084ABC 2.412705058144.2
085M99 2.2.112716248144.3
086BOA 0.58b12731521144.5
087PPMX 0.0712734291144.5
088ShipInBottle 1.0 b1712737327144.5
089CHILE 0.512740959144.6
090WINIMP 1.2112744029144.6
091QAZAR 0.0pre512749849144.7
092DST 0.91b12768962144.9
093CABARC 1.00.010612801613145.3
094X1 0.95a12852652145.8
095IMP 1.1212853035145.8
096WinTurtle 1.6.012924324146.6
097DGCA 1.1012962432147.1
098BIX 1.00b712966506147.1
099MAR12974461147.2
100GCA 0.9k13033460147.9
101ACE 2.613052581148.1
102WinACE 2.6913052829148.1
103DCGA b813092688148.6
104QUANTUM 0.9713100146148.6
105TURTLE 0.0713110829148.8
106BICOM 1.0113121337148.9
107BSA 2.0013144755149.1
108ARHANGEL 1.4013169564149.4
109DACT 0.8.4213172766149.5
110NNTC13237834150.2
111OrangeArchiver 1.0513265652150.5
112BWTZIP13268400150.6
113BZIP 0.2113348813151.5
114BZIP2 1.0.513379118151.8
115CSC 3.2a613379276151.8
116PSA 0.91a13414981152.2
117LZXQ 0.413487115153.0
118PAR 2.0013491351153.1
119EXP 1.013502811153.2
120PAC 17apr200413515568153.4
121PACKET 0.91a13539533153.6
122RZIP 2.113555841153.8
123Tornado 0.4a13595881154.3
124LZDS v2.113625557154.6
125SR3a13643600154.8
126BZP 0.313696214155.4
127LGHA 1.1g13732210155.8
128HA 0.999b13733651155.8
129TRANSFORM 1.0213750394156.0
130ASD 0.2.013774725156.3
131XPv513836270157.0
132Comprox 0.3.013857760157.2
133LZAP 0.20.0b13857867157.2
134JAR 1.0213862146157.3
135EXTREME 1.0613865425157.3
13612Ghosts 7.013873112157.4
137AI 1.113890894157.6
138CODEC 3.2113906106157.8
139PPMT 0.113914385157.9
140ARI 2.213997766158.8
141THOR 0.9614120888160.2
142LZ2A14144400160.5
143KZIP 14-APR-200714280477162.0
144YAC 1.0214325263162.5
145DCA 1.0.1b14337289162.7
146SEMONE 0.614393960163.3
147SLUG X14456960164.0
148UC II v3.05b14551136165.1
149DeepFreezer 1.0614555458165.2
150LHA32 1.88.3.1414564115165.3
151LHARK 0.4d14571843165.3
152DLC 0.6.114672708166.5
153BJWFLATE 1.5414753711167.4
154RDMC 0.06c14783483167.7
155SYMBRA 0.214843240168.4
156HIT 2.1014845671168.4
157ZET 0.10b14865875168.7
158LIMIT 1.214868331168.7
159PKZIP 2.5014868451168.7
160ZAP32 0.15.0b14871952168.7
161AMG 2.214872858168.8
162HuffComp 1.314878310168.8
163DZIP 2.9014928634169.4
164vuZIP 1.814930430169.4
165EAZEL 1.014930765169.4
166LHA 2.6714939058169.5
167GZIP 1.3.514948376169.6
168ZIP 2.214948761169.6
169AIN 2.3214964305169.8
170ESP 1.9214970505169.9
171WINZIP 8.014994002170.1
172JCALG1 5.3215047102170.7
173ARJ 2.8515059627170.9
174Crush 0.0115069847171.0
175WIN-GZ 1.215087157171.2
176File2Pack 2.015088358171.2
177 BCArchive 1.08.715147600171.9
178RAX 1.0215184339172.3
179Windows XP built-in15283775173.4
180HAP 3.0615290714173.5
181BVI 1.7015461581175.4
182aPLib 0.4315567219176.6
183ALZip 7.015619459177.2
184XPA 1.0.215623505177.3
185LZA 1.0115701975178.2
186Chaos Comp 3.015830671179.6
187LCSSR 0.215842346179.8
188Etincelle RC215933095180.8
189SAR 1.015989247181.4
190AR 1.015989247181.4
191ZOO 2.115990584181.4
192Archiver 1.016028708181.9
193ULZ 0.0.216030541181.9
194BAR 1.1.216038084182.0
195HiP beta 116113154182.8
196HYPER 2.516469406186.9
197LZOP 1.02rc116605317188.4
198LZC 0.0816895049191.7
199Zhuff 0.217293784196.2
200ERI 5.1fre17317756196.5
201QuickLZ 1.40b917325712196.6
202ARX 1.017360261197.0
203Secura 1.717748104201.4
204CA-ZIP 3.417897153203.1
205CODER 1.118150341205.9
206QPress 0.38b18394649208.7
207BigCrunch 0.4a119003791215.6
208BriefLZ 1.0419150759217.3
209LCW 0.219503456221.3
210LZ 1.020402392231.5
211SRANK 1.021222999240.8
212LZP2 0.7d21395263242.8
213LZBW1 0.823322360264.6
214LZRW124878204282.3
215Shindlet36659263416.0
216SHcodec 1.0.138815350440.4

The complete ranking list can be downloaded (right click, save link as in Firefox, or save target as in IE).

- Scoring system: The compressed size of all 10 tests per program are added up.
The program with the lowest total is considered the best program.


Q. Now, what is the best compression program?.
A. That depends on your definition of 'best'. Globally speaking there are three kind of compressors / archivers.

First we have the group of experimental compressors, designed with only one thing in mind, getting the best possible compression regardless the time it takes and the amount of memory it uses. Almost all programs in this group are command line (dos/win32) programs, and some of them can only compress one file at a time (remember user friendliness wasn't a real consideration). The best programs in this group, i.e. the ones getting the best possible compression, are WinRK (using PWCM mode), PAQ8, PAQAR, Durilca and SLIM. At the moment WinRK (PWCM) and PAQ8 are the two best compressors of the five, but they are also much slower then Durilca and SLIM (to give you an idea; PAQAR in -8 mode does about 5-7 KB/s on a slightly overclocked AMD Barton 2800+). If you have files with embedded images (e.g. Word DOC files) use PAQ8, it will recognize them and separately compress them, boosting compression significantly. Also note all mentioned programs with the exception of WinRK are free of cost.

Secondly we have a group which tries to combine the good compression of group one with some degree of user friendliness. Compression is still very good, speed is reasonable, but the compression formats used are often not widely spread. Best programs is this group are UHarc (with WinUHA gui) and WinRK.

The last group is the 'everyday use' archiver. This are GUI driven programs with good compression, but without sacrificing speed and without excessive memory usage like the group 1 and 2 programs. Apart from the GUI it is important the program is fast, easy to use, and supports many different archive formats (or is a standard itself). The best archivers in this group are WinRar and 7-Zip. Maybe WinZip 10 (using PPMd-mode) has a place here too.


Q. Why didn't you include program xxxxx ?
A. I try to keep the comparison benchmarks fair and transparent by only listing one single version of a program (the one with the best compression). For that reason I don't include clones of a program (like PAQ clones WinUDA, Emilcont and KGB), 'light' versions (like Durilca Light) and programs without an own compression format (like PowerArchiver, Total Commander, ZipGenius, TUGzip). If you still think I missed a program please let me know!. You can mail me (Werner Bergmans) at 'info _AT sign_ thissitename dot com' (or have a look at the front page to see this email address spelled out...), or -prefferably- post your suggestions in the Maximum Compression Guestbook. PS. Most of these clones and light versions are included in the multiple file compression (MFC) test.


Q. Why don't you add compression times?
A. As the sites' name already indicates, I'm looking for the maximum compression ratio. The time it takes is irrelevant (to me), getting the best switch combinations takes by far the most time... I agree time is useful to know if you are considering an 'every day use' compressor, but that's not what the site is about.

Apart from the fact it takes a lot of time to do all measurements there is a bigger problem. I try to find the best combination of switches to yield maximum compression (and this really can be an exotic combination for programs like 7-Zip, Durilca etc). That particular combination can lead to say 40.00% compression with a compression speed of say 20 Kb/s. A different set of switches can have 39.99% compression with a speed of 500 Kb/s. Is it fair to list this 20 Kb/s as, 99.9% of the people will not use the switches I used?. I think speed comparisions are only fair if you let the program use it's default settings. The multiple file compression test (MFC) does include (de)compression times are more 'regular' switches / options are used there.

Next Test Home Previous Test

©2003-2011 MaximumCompression (lossless data compression software benchmarks)
email