Powhertz Archiver Test 2000

POWHERTZ ARCHIVER TEST 2000 -
THE RESULTS PAGE

The 27 archivers of this test are compared on 7 different samples:

one text file (30 459 bytes)
one Word 97 binary file (107 008 bytes)
one MPEG video file (2 267 140 bytes)
one MPEG3 audio file, 192 kbps-encoded (5 889 567 bytes)
one Bitmap image (24-bit color) (1 140 056 bytes)
one WAVE audio file (876 180 bytes)
a complete game, with subdirectories, hasardly chosen on Powhertz' computer (the demo version of The House of the Dead, 75 887 368 bytes total, including a very big textures file, a big executable, many sound files, etc...)

Here are the results. The first row is the number of bytes of the compressed file (the lower, the better), the other numbers are the compress time and the uncompress time, in seconds or in minutes depending on the scale. Note that, when possible, the archivers are always tested using their optimal compression method (look at the test notes for exceptions).

Archiver	Text	Word 97	MPEG	MPEG3 (.MP3)	Bitmap (.BMP)	WAVE	Game (THOTD)
777 v0.04 Beta 1	11 274	14 716	2 049 555	5 751 626	546 992	858 056	Abandonned - unfunctional recurse
	2 s. / 2 s.	14 s. / 4 s.	10:54 / 9:47	31:02 / 28:28	2:14 / 1:11	4:42 / 4:19	---
ACB 2.00c	11 176	16 514	2 047 814	5 789 256	768 896	874 591	12 260 255
	1 s. / 2 s.	1 s. / 2 s.	2:49 / 2:50	5:05 / 5:17	1:13 / 1:14	41 s. / 41 s.	? / 62:00+
ARC 6.02	17 566	46 908	2 210 993	5 889 598	1 178 123	876 211	20 723 038
	0 s. / 0 s.	0 s. / 0 s.	4 s. / 1 s.	11 s. / 2 s.	2 s. / 1 s.	2 s. / 0 s.	1:17 / 37 s.
ARHANGEL 1.40	10 819	"Not enough memory" error	2 096 314	5 889 622	853 046	826 684	14 085 722
	1 s. / 0 s.	---	11 s. / 7 s.	26 s. / 1 s.	6 s. / 3 s.	34 s. / 34 s.	7:16 / 2:57
ARJ 2.75	12 904	18 922	2 093 672	5 810 675	859 383	871 536	14 407 786
	0 s. / 0 s.	1 s. / 0 s.	5 s. / 1 s.	13 s. / 4 s.	3 s. / 1 s.	2 s. / 1 s.	2:30 / 27 s.
ARJ32 3.08	12 908	18 873	2 093 516	5 810 578	859 173	871 536	14 380 114
	0 s. / 0 s.	1 s. / 0 s.	6 s. / 2 s.	15 s. / 6 s.	3 s. / 1 s.	3 s. / 1 s.	3:11 / 29 s.
BOA Constrictor Archiver 0.58b	10 439	17 210	2 037 952	5 741 556	531 953	878 176	"Not enough memory" error
	1 s. / 1 s.	2 s. / 2 s.	5:33 / 6:22	16:10 / 18:32	55 s. / 59 s.	2:15 / 1 s.	---
CuteZip 1.0	12 893	18 319	2 092 071	5 798 128	855 235	870 987	14 340 293
	1 s. / 0 s.	1 s. / 1 s.	8 s. / 1 s.	16 s. / 4 s.	5 s. / 1 s.	3 s. / 0 s.	9:00 / 25 s.
Free Zip 2.2.2	12 867	18 293	2 092 045	5 798 102	855 209	870 961	Makes the application crash!
	1 s. / 0 s.	1 s. / 1 s.	6 s. / 4 s.	16 s. / 11 s.	5 s. / 3 s.	2 s. / 2 s.	---
JAR 1.02 (JAR32)	11 505	17 257	2 081 258	5 795 893	839 174	874 314	13 652 598
	1 s. / 0 s.	2 s. / 0 s.	11 s. / 3 s.	27 s. / 7 s.	9 s. / 2 s.	5 s. / 2 s.	9:30 / 56 s.
LHA 2.55e	13 202	26 916	2 101 557	5 822 952	868 053	871 278	14 921 732
	0 s. / 0 s.	0 s. / 0 s.	6 s. / 1 s.	15 s. / 4 s.	5 s. / 1 s.	2 s. / 0 s.	3:38 / 29 s.
lzop 1.00w	14 686	19 337	2 132 986	5 853 520	958 383	876 282	Abandonned - no recurse
	0 s. / 0 s.	1 s. / 0 s.	10 s. / 1 s.	27 s. / 1 s.	8 s. / 0 s.	5 s. / 0 s.	---
PKZip 2.50 (DOS 32 bits)	12 838	18 485	2 091 989	5 803 887	851 060	876 119	14 243 651
	0 s. / 0 s.	0 s. / 0 s.	4 s. / 1 s.	10 s. / 3 s.	2 s. / 1 s.	2 s. / 1 s.	2:02 / 21 s.
PKZip 2.70 (Windows 32 bits)	12 838	18 485	2 091 985	5 803 885	851 048	876 119	14 243 670
	0 s. / 0 s.	0 s. / 0 s.	4 s. / 1 s.	10 s. / 2 s.	2 s. / 0 s.	2 s. / 0 s.	2:05 / 23 s.
PPMD vF	10 455	20 479	2 065 104	5 844 034	758 082	888 301	Abandonned - no recurse
	0 s. / 0 s.	1 s. / 1 s.	46 s. / 52 s.	2:11 / 2:35	13 s. / 14 s.	20 s. / 24 s.	---
RAR 2.50 (DOS)	12 790	17 693	2 084 710	5 804 076	838 332	786 508	13 855 954
	1 s. / 0 s.	1 s. / 0 s.	11 s. / 2 s.	21 s. / 6 s.	4 s. / 1 s.	3 s. / 3 s.	5:07 / 31 s.
RAR32 2.60 (DOS 32 bits)	12 713	16 796	2 079 905	5 794 783	827 424	786 463	11 068 870
	0 s. / 0 s.	0 s. / 0 s.	13 s. / 2 s.	33 s. / 5 s.	7 s. / 1 s.	4 s. / 2 s.	5:45 / 27 s.
RK 1.02 alpha 5	10 604	15 980	2 093 256	5 832 120	520 556	889 972	13 505 504
	1 s. / 1 s.	5 s. / 5 s.	27 s. / 29 s.	58 s. / 55 s.	56 s. / 57 s.	15 s. / 18 s.	3:19 / 2:43
RKIVE 1.92 Beta 1	10 485	16 178	2 113 451	5 962 737	414 833	949 523	12 451 386*
	3 s. / 2 s.	1 s. / 1 s.	5:46 / 5:31	16:51 / 17:12	9 s. / 10 s.	2:26 / 2:31	29:45 / ---
SuperZip 2.0	12 901	18 352	2 092 085	5 798 136	855 542	870 995	Abandonned - no recurse
	0 s. / 0 s.	1 s. / 0 s.	4 s. / 1 s.	11 s. / 4 s.	3 s. / 1 s.	2 s. / 0 s.	---
AIP-NL UltraCompressor 2.37b	12 402	18 094	2 084 270	5 801 100	631 842	843 630	13 906 998
	1 s. / 0 s.	1 s. / 1 s.	12 s. / 3 s.	32 s. / 7 s.	31 s. / 1 s.	13 s. / 1 s.	11:56 / 36 s.
UFA 0.04 Beta 1	11 274	15 277	2 049 690	5 889 661	510 017	868 193	11 860 866
	1 s. / 1 s.	8 s. / 0 s.	1:32 / 1:38	3 s. / 2s.	1:04 / 11 s.	20 s. / 11 s.	58:21 / 3:44
WinACE 1.31	12 817	16 659	2 083 282	5 804 294	833 737	872 744	13 303 396
	1 s. / 1 s.	2 s. / 2s.	10 s. / 3 s.	27 s. / 6 s.	6 s. / 2 s.	4 s. / 2 s.	3:27 / 37 s.
WinIMP 1.2	12 709	17 102	2 070 709	5 778 840	827 357	870 830	13 539 931
	0 s. / 0 s.	1 s. / 0 s.	7 s. / 1 s.	18 s. / 2 s.	4 s. / 0 s.	2 s. / 0 s.	7:21 / 22 s.
WinZip 8.0	12 867	18 293	2 092 045	5 798 102	855 209	870 961	14 342 013
	0 s. / 0 s.	1 s. / 0 s.	5 s. / 0 s.	12 s. / 3 s.	3 s. / 0 s.	2 s. / 0 s.	5:52 / 25 s.
WinRAR 2.71	12 651	16 762	2 073 826	5 788 323	823 035	871 497	13 298 835
	1 s. / 0 s.	2 s. / 0 s.	19 s. / 2 s.	59 s. / 6 s.	9 s. / 2 s.	6 s. / 1 s.	8:01 / 34 s.
zoo 2.1	13 333*	27 048	2 101 689	5 823 083	868 184	871 409	Abandonned - no recurse
	0 s. / 0 s.	0 s. / 0 s.	9 s. / 2 s.	21 s. / 5 s.	7 s. / 1 s.	3 s. / 1 s.	---
Archiver	Text	Word 97	MPEG	MPEG3 (.MP3)	Bitmap (.BMP)	WAVE	Game (THOTD)

Key: black = DOS command line blue = DOS Interface, textual but presented graphically green= GUI Interface red = category champ
* "CRC Failed"-type error while tying to decompress the created .RKV archive and "I/O Error"-type error while trying to decompress the resulting .ZOO archive. This means we shouldn't take too seriously those results.

THE TEST NOTES - VARIOUS NOTES ON EACH ARCHIVER

777 0.04 Beta 1

Different algorithms available. For optimal compression, the "Text" algo. has been used on the text file, the "Bitmap 24-bit" algo. has been used on the Bitmap file, and the option "-mg" (for "Method = Guess Best") has been used for the 4 other tests. This explains the strange-looking compress & decompress times much higher for the WAVE than for the Bitmap while the original size of the Bitmap file was greater.
There is a recurse option (to include in the archive the subdirectories and the files they contain), but it doesn't seem to work. The "-mx" option (for "Maximize compression") doesn't seem to work either.
777 is an experimental version of UFA, from the same author.
Supports the long filenames.

ACB 2.00c

Good: supports volumes.
Bad: no choice of compression level; there is only one mode, which compresses very well but which is also pretty slow.
Supports the long filenames.

ARC 6.02

No choice of compression level; ARC is automatically the fastest but also the one with the lowest compression.
Let's mention that the documentation of ARC specifies that it's on spreadsheets (MS Excel, Lotus 1-2-3, Quattro Pro, etc...) that ARC is at its best, but there is no file of that type in this test.
Doesn't support the long filenames.

ARHANGEL 1.40

Does not really have a choice of compression level, but there are different algorithms available. For this test, the "multimédia" algo. has been used on the WAVE file but the standard algorithm has been used for all the rest.
According to its documentation, it's on 24-bit TIFF image files that ARHANGEL is the most efficient and there aren't any in this test.
Supports the long filenames (it's the difference between versions 1.33 and 1.4!).

ARJ 2.75

There is a maximal compression option ("-jm1"), but it haven't been used for this test because I just forgotten.
Doesn't support the long filenames.

ARJ32 3.08

Used the maximal compression option ("-jm1") for this test.
Doesn't work at all under MS-DOS 6.x, despite a command-line interface, it is a real Win32-specific application..
Built on the same foundations than ARJ, but written for Win32. The compression ratios are supposed to be very similar to those obtained with ARJ for DOS. The only true differences are the support of the VFAT filesystem extension, the support of Windows wildcards (* accepted instead of having to specify *.*) and a little bit lower speed.
Supports the long filenames.

BOA Constrictor Archiver 0.58b

Doesn't really work with compression levels but rather use the scheme of a number of MB of extended memory to use (from 1 to 15). The more megabytes, the better compression ratio but also the lower processing and the more megabytes needed to decompress the archive later. For this test, the maximum of 15 MB has been used for each, except the game because even 1 MB was too much!
Also used the "Solid" option for text file and the Word file since the documentation says that this option gives a better compression ratio on small files.
Contrary to what we may think, the standard extension for an archive is not .BOA, but rather .B58. The documentation mentions that when the project will be completed (version 1.0), the standard extention will become .BOA.
Supports the long filenames.

GlobalScape CuteZip 1.0

Characteristic to also support the compression of CD audio tracks to MPEG3 format.
Characteristic to do its archiving/compression job first and to only save the file after clicking on the Save button. For this test, the chrono times are the addition of the archiving time with the file-saving time.
Supports the long filenames.

Free Zip! 2.2.2

Free, but unfortunately too buggy to be considered as a true alternative to WinZip and PKZip which are much too expensive. While using it only a few minutes, already the bugs were showing and one of them was even making the application crash every time I was trying to archive the complete game.
Cute toolbar, but unfortunately, to keep the program free, it has some banners.
Supports the long filenames.

JAR 1.02

JAR is composed of 2 different executables files, JAR16.EXE and JAR32.EXE. For the test, the 32-bit version has been used.
Choice of various different algorithms. As it is mentionned in the documentation that the "-m4" option gives a greater compression ratio on most files of 2 MB and up in size, this option was used fot the tests on the MPEG, the MPEG3 and the complete game while the "-m3" option was used for the 4 other tests.
Supports the long filenames.

LHA 2.55e

No choice of compression level; has only one algorithm, pretty weak but yet pretty fast.
For those of you who wouldn't know, this archiver, popular in the early '90s, creates archives with .LZH as the file extension, not .LHA as some could think. LHA is only the name of the program, there aren't .LHA files.
Originally (and still) conceived to work on any PC, even an old 8086 (the ancestor of the 80286 processor, released in the '80s).
I read that LHA was at version 2.67, but the most recent version I've been able to find for download is 2.55e.
Supports the long filenames.

lzop 1.00w

The default extension of the archives is .LZO, not .LZP as we may guess.
Many compression levels available. For this test, the "-9" option (best compression) has been used.
Offered under the famous GPL license, which means it is 100% free and even open-source.
Supports the long filenames.

NetZip Classic 7.5.1.85

Strictly refused to install on my Windows 95B system! The installation program was always crashing at the same place near the end of the installation process. I've never been able to complete the installation, so the program has just been removed from the candidates.

PKZip 2.50 (DOS 32 bits)

All of the options changed since the legendary version 2.04, but the new options are all equivalent to old ones and conversion table is provided in the documentation. Of course, the "-max" option (equivalent to the "-ex" of version 2.04) has been used for all of the tests.
Supports the long filenames.

PKZip 2.70 (Windows 32 bits)

Only 2 drawbacks: it doesn't support the "drag & drop" for extraction (although it well supports d&d for adding files to an archive) and the pub banners server it uses installs itself under the "Run" key of the Windows registry, which means that it implicitely starts up at every Windows boot-up and that it is, like, always in memory.
Supports the long filenames.

PPMD vF

Also known as "Fast PPM compressor for textual data". As its name suggests, this archiver mostly optimized for text files compression.
Works with a number of MB of RAM to use. For this test, 16 MB of RAM have been specified.
Option of optimisation for compression of executables files on Genuine Intel Pentium processors. This option has been used for the tests since the machine used for the test is driven by an Intel Pentium MMX processor.
Not only free, PPMD is also open-source.
Supports the long filenames.

RAR 2.50 (DOS)

One of the only 2 DOS archivers DOS to feature a not-command-line-only interface, RAR's interface is better than the one of its rival (UltraCompressor).
The "Best compression" option has, of course, been used for these tests, but the "multimedia compression" option has also been used on the WAVE file and the MPEG video.
Doesn't support the long filenames.

RAR32 2.60 (DOS 32 bits)

In contrary to ARJ vs. ARJ32 or JAR16 vs. JAR32, RAR and RAR32 are pretty much different. They aren't born exactly on the same the same base programming and the 32-bit version is in command-line, in contrary to the 16-bit DOS version.
Supports the long filenames.

RK High Performance Archiver v1.02 build 5 (alpha)

Pretty light documentation.
"multimédia" method used for the MPEG file, "WAVE compression" method used for the WAVE file, standard method for everythig else.
Option for maximum compression ("-mx") also used in each case except for the MPEG3 (the archiving was only 29% done after one hour, so because of the excessive time, I restarted at normal compression) and for the complete fame (only 3% done after 45 minutes; so I restarted at normal compression).
RK was by far the archiver doing the most hard disk inputs/outputs in this test. In opposite to other slow archivers like 777 and BOA which mostly seemed to busy up the processor, RK was constantly working on the disk, which makes me believe that its performances could have been much better with 64 MB of RAM and a faster hard drive.
Supports the long filenames.

RKIVE High Performance Archiver v1.92beta1

Much better documentation than RK of the same author.
The standard archive extension is .RKV .
Has a "-mt3" method optimized for text and used on the text file in this test.
The other main method is the "Binaries" method with different levels of compression. The "-mb3" option has then been selected for all of the tests except for the text (rather used text method).
Supports the long filenames.

SuperZip 2.0

Doesn't seem to support the recurse of subdirectories.
Supports the long filenames.

AIP-NL UltraCompressor 2.37b

Not bad RAR-like interface, except for the decompression which seems impossible from this interface. Characteristic: for the same executable file, if you only type the filename at the DOS prompt, you end up in small menu, if you specify the VU switch, you go directly into the tex-mode graphical interface, and if you specify other switches, you use the program in command-line!
For this test, the "S-TIGHT" compression level has been used.
Supports the long filenames. Indeed, the version 2.37b is the first to do so..

UFA 0.04 Beta 1

Differents algorithms available. For an optimal compression, the "Text" algo. has been used on the text file, the "Bitmap 24-bit" algo. has been used on the Bitmap file, and the "-mg" option (pour "Method = Guess Best") has been used for the 4 other tests. This explains the strange-looking compress and decompress times much greater for the WAVE file than for the Bitmap while the Bitmap's original size was higher.
Is this intelligence or the exact opposite? While using the option to guess the best algorithm, UFA concludes that it is even better not to spend any time to compress the MPEG3 file! This explains, of course, while the compression and decompression times are so low for this test.
Unlike its experimental cousin 777, UFA has not only a choice of algorithm, but also a "Maximize compression" option that was used in all of the tests except for the complete game (with maximum compression, it was only 20% done in 1::10...).
Supports the long filenames.

WinACE 1.31

Fat, overloaded and hard to masterize interface.
Supports the long filenames.

WinIMP 1.2

Supports the long filenames.

WinZip 8.0

Tests done with the classic interface (shouldn't really have too much influence on performances, but if one of the 2 interfaces available is better than the other, it is this one).
Interface sometimes well-conceived, sometimes pretty bad. For example, the "add files" dialog just fits on a screen at a resolution of 640x480 while there is waste of space on the dialog. Also, the "drag & drop" of a directory is very confusing, just stupidly-conceived.
Supports the long filenames.

WinRAR 2.71

Huge (but not necessarly cute) icons; completely unuseful waste of space.
Supports the long filenames.

zoo 2.1

Undoubtly the most complex of use in the 27 archivers of this test.
No choice of compression level, but there is a "h" switch for "high performance".
Doesn't support the long filenames.

ICONS
I know, completely unuseful you'll say, and you're right, but it is still the icon you'll see everywhere on your Windows system. Here are the icons I found the cutest, in order:
1- Free Zip
2- PKZip
3- WinACE
4- WinRAR
5- WinZip
6- CuteZip (in fact, probably cutest than WinZip and WinRAR, but I just didn't understand its meaning!)
7- WinIMP
8- SuperZip (very ugly!)

ANALYSIS
The text file really took place to an interesting battle. Each bit is important with a file that small, and the confrontation was very solid between the various lesser-known small archivers. BOA Constrictor just went past PPMD at the finish line, PPMD yet being specialized in text files: a big point for BOA. But check how much RKIVE and RK finish close, and even ARHANGEL would deserve to be mentionned here. On the other end, only ARC and lzop have been shown particularly weak.

For the binary file in Word 97 format, 777 ranks before anything else. Despite good performances in general, it is the only championship that 777 obtained in this test. And guess who comes second on the Word file? UFA, of which 777 is the experimental version! However, notice the great performance from RK that have been able to rank in the only 3 archivers doing below 16 000 bytes. Now, for the losers, let's say that ARC has largely won the award for the most atrocious, but the other ancestral archivers that are zoo and LHA also had a hard time. Note that PPMD also did above 20 000 bytes.

Came to the MPEG file, the battle seems a little bit less fierce. BOA Constrictor wins an almost-easy victory while, surprise, ACB takes the silver medal. Except for ARC (the worst of all again), the weakest ones are not too far of the pack in this heat.

Yes, of course, an MPEG3 file is already compressed and it is rarely useful to use an archiver over that, but it's always interesting to see which ones will do the extra work when it's more tight. The answer, once again, is BOA Constrictor, a 3rd championship in 4 tests. It is definitely an impressive one, but it is not alone because 777 is also much superior to the average and even compared to its cousin UFA. But the most surprising with the MP3 is certainly to see who end up at the bottom of the chart. RKIVE is the worst, followed by UFA and ARHANGEL! Against all expectations, ARC is only the 4th worst this time! Notice that those 4 archivers created an archive bigger in size than the original file...

I don't know what you think, but me, the battle I liked the most was the one for the Bitmap file compression championship. RKIVE was, like, all alone at the finish line, showing a world-record-setter performance! Wow! Its compression ratio is absolutely amazing (more than 63%!) and its times aren't bad at all! RKIVE being just an unreachable level, we can also boast UFA and RK which performances were absolutely great too. Finishing 4th and 5th, BOA Constrictor and 777 also made an archive under 600 000 bytes while the original file was above 1,1 MB. Great competition. Notice also the performance of UltraCompressor that ranks easily first amongst the archivers who did the first Powhertz test in 1998. At the opposite, ARC and lzop are the only ones to have put up an archive over 900 000 in size, very bad performances.

The competition slows down a little bit with the WAVE file. Except for RAR32 and RAR that, surprisingly, easily take the 1st and 2nd places, there is not a very fierce battle; the compression ratios are pretty much standard... except maybe for RKIVE and RK that are pretty much deceiving with their archive much bigger than the original file...

And finally, the most important which is of course the complete game because it is the only test evaluate multiple-files archives (with directories) and because of the size of the thing to archive, the size and time differentials are much greater. How spectacular, the winner of this test seems to come from nowhere and left me completely speechless. Before doing the test, I would have think possible that neither UltraCompressor, neither IMP wins this time, but as long as the new winner is a new one like BOA Constrictor or 777! Believe it or not, those 2 unfortunately haven't even been able to pass the test, along with Free Zip, lzop, PPMD, RKIVE and zoo which were all not even able to qualify. The winner is simply RAR32 which, surprisingly, has a neat step over RAR for DOS and WinRAR, also from RARSoft. UFA and ACB finishes up excellent 2nd and 3rd on compression, but with compression and decompression that just doesn't compare to those of RAR32. Really, in this test, RAR32 is the one and only, the unquestionable master. Except for the 7 unqualified (no directory recurse or the recurse option doesn't work, memory error even at the lowest compression ratio available, corrupted archive or application crash), ARC finishes last fairly easily, but the popular formats ARJ, LHA and ZIP (with any program) look nevertheless stupid with their archives in the 14 MB... I also need to my deception over ARHANGEL which joins the club of the weaks with an archive of 14 085 722 octets.

CONCLUSIONS
The simple way to conclude this test would be: "Hurrah for RAR32!". It takes the honors for the highest compression at 2 of the 7 tests including the most important, has shown a better compression ratio than the ZIP format (no matter the ZIP program used) in every test and did this with decent times. Congratulations, the winner purely and simply.

The longer way to conclude would of course be to say that all depends on the user. On a PC equiped with 64 MB of RAM and up, BOA Constrictor may be an excellent choice... if the compression ratio is an higher concern than the time. If the compression time if not your priority, I wouldn't hesitate to also recommand you UFA which is very solid and, again, still in beta. It is very promising for the final version... 777 is definitely another archiver to check out in the future, but for now I would prefer its cousin UFA because 777 isn't a completed project (some functions such as the recurse aren't functional for now). Nevertheless having a good word for ACB, surprising also, but so slow, as much for decompression than for compression.

Those recommandations are of course for general use... If you use to compress lots of text, then you'll maybe want to make your own tests between BOA Constrictor and PPMD, those two being very good candidates. If you are a webmaster, it is possible that you have, sometimes, to compress MPEG or MPEG3 files for faster download times and to unable your server to recognize an unauthorized file extension. In this case, BOA Constrictor is undoubtly what you need. If, for any reason, you compress many Bitmap images (it can happen that you want to copy a Windows wallpaper, which can be very big, and that you must make it fit on only one disk, for example), don't ask and download RKIVE right now (being aware that it can, sometimes, create corrupt archives) or, if you don't want to start collecting archivers, use UFA, which is also very efficient on other kinds of files.

Notice, all of the archivers that I particularly recommand are in command line (RAR32, BOA Constrictor, UFA). I know for sure that in year 2000, most people limit as much as possible their use of MS-DOS command lines. This is why I'll also recommand WinRAR, mainly to be the GUI ("Graphical User Interface") archiver with the highest compression rates, and PKZip for Win32 to be incredibly fast and to be native to the most widely-used archive format, the ZIP. WinIMP is also an excellent overall option, especially since it became free (version 1.2).

If I have many archivers to recommand, I also have many ones to advice against. The first one is certainly NetZip Classic, which problems starts with the installation! The second one would be, without too much hesitation, Free Zip, filled with nasty bugs. This version is led as "2.2.2", but should be considered as a "0.2.2" at most. The 3rd worst would be ARC, with the worst compression level at 5 of the 7 tests, sometimes even far behind the 26th. Follow up would be zoo, another old archiver that isn't competitive at all anymore, does not even support recurse and having put up a corrupt archive on the text file. I would never recommand either CuteZip, which takes much more time to do its compression jobs than its main rivals (WinZip, PKZip for Win32, SuperZip) for not better compression, and lzop, of relatively bad compression and without recurse support. RKIVE, 777, ARHANGEL, BOA Constrictor and SuperZip are all to recommand with precaution for having known some problems throughout the test (no directory recurse, not enough memory to complete an archival, corrupt archive). I would also tend to advice against WinACE for its fat, overloaded interface and its bad layout.

Click here to download an archiver of the test or to know the official Web site address of the archivers

Powhertz
2000/10/28

Retour à la page principale de Powhertz / Back to Powhertz main page

Tests done on an Intel Pentium MMX computer clocked at 233 MHz, equiped with 32 MB of RAM and operated by Microsoft Windows 95B. The partition on which all these tests were made have been reformatted just before the test and only contained Windows 95B, the 28 archivers to test and the "guinea-pig" files.