MONSTER OF COMPRESSION - New Benchmark -

[Nania Francesco Antonio]

MONSTER OF COMPRESSION
23.04.2008
- Added Fpaq0pv4B
- Added 4X4 v.0.2
- Updated Rings 1.5b
- Updated CMM4 v.01e
- Updated BalZ 1.03
http://www.winturtle.netsons.org/MOC/MOC.htm

[Christian]

Thanks for adding Slug, too. As a note, Slug 1.2+ does not use LZ77 anymore. It uses Huffman and a distant variation of ROLZ.

[toffer]

Thanks!

I'm curious about the timings. This version shouldn't be slower than 0.1d (i got no speed change on my testset). Or did you test other compressor options? If you would be so kind, could you please test the options 43?

[Nania Francesco Antonio]

Ok Christian! Corrected HUF+ROLZ !
I have not succeeded in making the DOWNLOAD of RZM and CCM!
CMM4 is High level compressor!;)

[Christian]

Quoting: Nania Francesco Antonio

I have not succeeded in making the DOWNLOAD of RZM and CCM!

Those damn Rapidshare cats, right? I'm sorry - next time I'll try to upload my programs to another hoster.
Anyway, there hasn't been much of a change - just a recompilation with GCC 4.3.0. So, you do not miss much.

[toffer]

Thanks again, but this still seems odd to me. Since cmm4 0.1d and e run at the same speed and both of them are faster than any previous cmm4 and 3 version. I would be very grateful for a short investigation.

[LovePimple]

Thanks Francesco! :)

Quoting: Nania Francesco Antonio

I have not succeeded in making the DOWNLOAD of RZM and CCM!

They are now available from my site.

http://www.geocities.com/lovepimple_mail/

[Nania Francesco Antonio]

Thanks LovePimple! In the next Update ! ritesterņ again cmm4 for the speed !

[toffer]

Thanks a lot, Nania!

[Raymond_NGhM]

I think www.RARhost.com better & simple than all,

one click>select .RAR file> upload...

No cats/no pusy cats/no dogs & no one ... ;)

[Raymond_NGhM]

Hi, Nania Francesco Antonio

Please try/add switch 5 for CCM(x) on MOC test,
then compare (De)Compression speed & ratio with previous switch,

[Nania Francesco Antonio]

Ok in the next release! HI!

[Nania Francesco Antonio]

MONSTER OF COMPRESSION 2007 Closed
Winner:
CCMX by Christian Martelock (size)
SLUG 1.23 by Christian Martelock (efficiency)
LZRW1 by Ross Williams (Compression Time)
QUICKLZ 1.40 beta 5 by Lasse Mikkel Reinhold (Decompression Time)
MONSTER OF COMPRESSION 2008 Incoming...............

[LovePimple]

Thanks Francesco! :) SO! CCMX is the MONSTER for 2007!

Congrats to all who won.

[Nania Francesco Antonio]

MONSTER OF COMPRESSION 2008
BY NANIA FRANCESCO ANTONIO

RELEASED 30.04.2008

NEW COMPETITION:
This pages ranks lossless data compression programs by the compressed size of the 30 types of file for 729 MB (765.008.542 byte) bytes:
BMP,TIFF,TGA,GIF, PNG, PSD, PGM, HTML, SYS, DIC,
DLL,DOC,EXE,HLP,JPG, LOG, MP3,WMA, PDF , SAVE,
TAR, TXT, VOB, WAV, XLS, AC3, AVI, ISO, BIN.

NEW RULES:
1) Max 1 GB of memory;
2) Time limit= 1.800 sec. (30 min.) for compression ;
3) No external dictionaries;
4) No errors!
5) EFFICIENCY = (POWER(2; ((size_X / size_TOP) - 1) / 0,1) * time_X)+(size X/500.000).
6) 4 Ranks – SIZE – COMP. TIME – DEC. TIME - EFFICIENCY.

[LovePimple]

IMHO those rules are very fair. :)

Let the competition begin!

[encode]

Quoting: LovePimple

Let the competition begin!

Let the Mortal Kombat begin!
(C) Shang Tsung

[Christian]

Hi Francesco!

Thanks a lot for MOC. Great that there is a new instance for 2008. Just some things I noticed.

-CCM is using a "(fast) context mixing" engine not "huffman + rolz".
-Can you please test RZM 0.07h - it uses less memory and is a bit faster.
-Can you please test Slug 1.24 - it's a bit faster and compression should be a tiny bit better.
-Compression time of Slug 1.23 shouldn't be ~50% more than decompression time - even the IO is exactly the same for compression/decompression.

[Christian]

Quoting: encode

Let the Mortal Kombat begin!
(C) Shang Tsung

Round one ... Fight

Maybe round two is more fitting. ;)

[Nania Francesco Antonio]

Have modified the description of CCM and CMMX!
Updated RZM 0.07h
Updated SLUG 1.24

[Raymond_NGhM]

Hi Nania

Please add/test RZM 0.07e
It faster than newer...
:)

[Raymond_NGhM]

I forget ...

Thanks for previous test. ;)

[Christian]

Quoting: Raymond_NGhM

Please add/test RZM 0.07e
It faster than newer...

On my C2D system 007h is always faster for compression and decompression - and uses 32M less memory. Taken from my development-tests (3 runs each):

Open-Office-Install (TAR) 
007e 274822k -> 65929k 222.8s 16.4s 
007h 274822k -> 65929k 216.0s 15.6s 

Results of this test: ~3% faster compression and ~5% faster decompression. Nonetheless, both versions are compatible. So, if 007e works better for you then there's no reason to use 007h (except the lower mem. usage).

[Raymond_NGhM]

-->if 007e works better for you then there's no reason to use 007h (except the lower mem. usage).


Yes Christian, at least on my single core,
Celeron 950Mhz/L1:32K/L2:128K (overclocked),
SSE support, w/512MB of SDRAM (CL:3-3-3-6)
;)

[Nania Francesco Antonio]

MONSTER OF COMPRESSION 2008
NEW UPDATE 02-05-2008
Added QuickLZ
Added CABARC
Added MNZIP
Updated WinRAR
Updated SLUG

[Simon Berger]

Are you sure with the compression time of Winuha? - It looks much too high. Also I would suggest to use Uharc instead because you can time much better and could use more switches. Winuha has also some bugs.

[Nania Francesco Antonio]

Winuha is very Slow compressor! Yes!

[encode]

I think UHARC should be tested! WinUHA is just a lame wrap around this command-line archiver... UHARC and ONLY! ;)

[Simon Berger]

Yes thats what I asked for ;)
I was hoping you would add decompression time to your efficiency formula or at least a second table. Is there any chance for? There are so many very good asynchron archivers which didn´t look as good as the should.

[Bulat Ziganshin]

specific of this test is a large amount of incompressible data. may be, uharc is too slow exactly on this type of data?

[Simon Berger]

Could be, but it is as slow as cmm4 and he tested the fastest mode. I don´t belive it ;)

[Christian]

Thanks for the update Francesco!

Quoting: Simon Berger

Could be, but it is as slow as cmm4 and he tested the fastest mode. I don´t belive it

Looking at the speed, ratio, memory usage and decompression time it must be a typo. I think -m3 was tested.

Btw., I'm wondering why there isn't any entry for CCM/CCMx using option 6? Memory usage is still easily under 1G.

Anyway, besides a few typos, you're doing a great job Francesco!

[LovePimple]

Thanks Francesco! :)

Any chance you could add the FastLZ (6pack) compile from my site?
http://www.geocities.com/lovepimple_mail/downloads/fastlz010.zip

If you decide to add them, please test both original and my experimental speed optimised (_opt) version. The speed optimised version does not compress quite as well, but compression speed is slightly faster on my machine. I'm interested to see how they compare with the other fast compressors on your MOC test.