MONSTER OF COMPRESSION - New Benchmark

[Nania Francesco]

Thanks endless for the help, if someone wants to signal me some compressor that are escaped me now thanks him/it since! Hi!

[osmanturan]

@nania:
First of all, you should leave TARred benchmark. This is unfair for statistical compressors and file extension based compressors (like FreeArc). AFAIK, MOC is started for finding best practical compressor. I don't use TAR before any compressor program for daily purposes. I think, most of people's habits like mine. I think, TARred benchmark only benefits for a compressor which have a large LZ dictionary and non-solid compressing capability.

@the others:
I think, CCM(x) surely use a LZ like submodel for dealing with highly redundant data. This also benefits for high speed compression. When LZ submodel is active, all other sub models seem inactive. Notice that, RZM seems slower and worser (I know RZM has almost no-filters. But, this comparasion is fair for me)

[Fallon]

5 (FIVE) RANKS

1- COMPRESSIONE SIZE (MONSTER OF COMPRESSION)
2- DECOMPRESSION EFFICIENCY (KING OF COMPRESSION)
3- COMPRESSION EFFICIENCY
4- COMPRESSION TIME
5- DECOMPRESSION TIME

Hi Francesco!
If you like, a suggestion to complete your 5 (five) MOC ranks.

1- COMPRESSION SIZE (MONSTER OF COMPRESSION)
2- DECOMPRESSION EFFICIENCY (KING OF COMPRESSION)
3- COMPRESSION EFFICIENCY (champion of compression)
4- COMPRESSION TIME (rocket of compression)
5- DECOMPRESSION TIME (the ripper of compression)

Of course other names are possible too.
Love the theme of your site!

[Nania Francesco]

06.07.2008

Added NanoZip
Added PPMD (re-tested)
Added Winzip 11.2
Added Squeez 5.61
Added Blizzard 0,24B
Updated Balz 1.14
Link:
http://www.winturtle.netsons.org/MOC/MOC.htm

[Sami]

Thanks a lot for testing Nania. I regret nz_optimum2 (-cO) is not tested, as I think it might be very efficient in your test because the decompression time is measured.

[Nania Francesco]

nz_optimum2 (-cO) tested in the next release! NanoZip is The Best compressor of the world for me!

[encode]

06.07.2008

Added NanoZip
Added PPMD (re-tested)
Added Winzip 11.2
Added Squeez 5.61
Added Blizzard 0,24B
Updated Balz 1.14
Link:
http://www.winturtle.netsons.org/MOC/MOC.htm

Thank you!
Note that BALZ v1.14 use just ~67 MB, for both compression and decompression!

[Christian]

Thank you for the update, Francesco!

Over the 'delightful' discussion with Sami I almost forgot to say thanks. Nonetheless, congrats to Sami for all the set records.

[Sami]

Thank you for the update, Francesco!

Over the 'delightful' discussion with Sami I almost forgot to say thanks. Nonetheless, congrats to Sami for all the set records.

Thanks again Chris. I'm eagerly waiting for the optimum2 test still, to see how it performs on the decompression efficiency.

[Zonder]

Hi Nania, why WinRK rolz3-fastest so slow on MOC?

on MFC and my tests WinRK is about 2-3x faster than in your tests...

[Nania Francesco]

@ Zonder , Hi!
I have made a will WinRK v.3.03 type of compression version nr.2 in the option fast, can sincerely tell you that almost all the programs of compression that use an arithmetic algorithm (ppmd,turtle,ccm.... and also winrk) in some types of file BIN (30 MB),MP3 (30 MB),WMA (30 MB),AVI (35 MB),WAVE (40MB),VOB (64 MB) etc....... you/they have slow down a lot the compression, not only but I have also noticed that those that use the system of search ROLZ, hardly a lot in these compressible files!

[IsName]

@nania:
First of all, you should leave TARred benchmark. This is unfair for statistical compressors and file extension based compressors (like FreeArc). AFAIK, MOC is started for finding best practical compressor. I don't use TAR before any compressor program for daily purposes. I think, most of people's habits like mine.

I agree with this! Nania, can you tell us why do you use a TARred benchmark? It's simpler for you? Or it's more fair in your opinion?

[Nania Francesco]

Unfortunately the translator from English to Italian doesn't allow me to understand at times well thing are asking me! Only a question but does someone perhaps believe that the benchmark MOC is based on an only file TAR? Because it deals with 765.008.542 byte 170 files in 30 folders!

[IsName]

Unfortunately the translator dall' English to Italian doesn't allow me to understand at times well thing are asking me! Only a question but does someone perhaps believe that the benchmark MOC is based on an only file TAR? Because it deals with 765.008.542 byte 170 files in 30 folders!

OK. In MaximumCompression benchmark, single file compressors tested with TAR, and multiple file compressors are tested without TAR (on files themselves).Thus TAR only used if needed. On MOC TAR is always used. FreeArc compresses TAR worse than files from TAR. Can you please unTAR testset and test FreeArc on files? I mean testset without TAR. Thanks.

[Bulat Ziganshin]

Unfortunately the translator from English to Italian doesn't allow me to understand at times well thing are asking me! Only a question but does someone perhaps believe that the benchmark MOC is based on an only file TAR? Because it deals with 765.008.542 byte 170 files in 30 folders!

sorry, but i can't believe you. fa 0.40 results was symmetric that proves that all files was compressed with ppmd algorithm. but fa 0.40 don't compress avi, jpg and other compressed files with ppmd. please show us the log of processing these data with fa 0.40

[Nania Francesco]

; MOC ARC TEST - SOURCE CODE -> IN PUREBASIC ;
InitKeyboard()
InitSprite()
Global Dim Path$(328000)
Global Dim nfile$(328000)
Global nufile,ndia,ndit
nudi:
ndit=1:ndia=1
inipath.s = Space(1000)
GetCurrentDirectory_(1000,@inipath)
Path$(ndia) = PathRequester("Please choose Source path", inipath)
ruspu2=Len(Path$(1))+1
For spu2=Len(Path$(1))-1 To 1 Step -1
If Mid(Path$(ndia),spu2,1)="\" : oldpath$=Mid(Path$(ndia),spu2+1,Len(Path$(ndia))-spu2):Goto lembo:EndIf
Next
lembo:
mepo:
nudivo=0
If ndia>ndit:Goto mepo2:EndIf
ExamineDirectory(0,Path$(ndia), "*.*")
lopas:
FileType = NextDirectoryEntry(0)
If Filetype>0 And DirectoryEntryType(0) = #PB_DirectoryEntry_File
nudivo+1
FileName$ = DirectoryEntryName(0)
If ReadFile(0,Path$(ndia)+FileName$)
nufile=nufile+1
nomef$=oldpath$+Mid(Path$(ndia),ruspu2,Len(Path$(n dia))-ruspu2+1)+FileName$
nfile$(nufile)=Path$(ndia)+FileName$
CloseFile(0)
EndIf
EndIf
If Filetype>0 And DirectoryEntryType(0)=#PB_DirectoryEntry_Directory
If DirectoryEntryName(0)<>"." And DirectoryEntryName(0)<>".." :ndit+1:Path$(ndit)=Path$(ndia)+DirectoryEntryName (0)+"\":EndIf
EndIf
If Filetype>0
Goto lopas
Else
ndia+1:Goto mepo
EndIf
mepo2:
If ReadFile(0,"c:\t.zip"):CloseFile(0) : DeleteFile("c:\t.zip"):EndIf
OpenWindow(0, 0, 0, 800, 1600, "IPER SQEEZE",#PB_Window_SystemMenu)
StartDrawing(WindowOutput(0))
For i=1 To nufile
DrawText(0,20+i*15," nome="+nfile$(i))
WindowEvent()
starti = GetTickCount_()
para$=" a -m6 "+" t.zip "+nfile$(i)
exe$="arc.exe"
result = RunProgram(exe$,para$,"c:\",#PB_Program_Wait| #PB_Program_Hide)
startf = GetTickCount_()
timenc=timenc+(startf-starti)
If ReadFile(0,nfile$(i)):lufa1=Lof(0):mefile =AllocateMemory(lufa1):ReadData(0,mefile,lufa1):Cl oseFile(0):EndIf
starti = GetTickCount_()
para1$=" e "+" t.zip"
exe1$="arc.exe"
result = RunProgram(exe1$,para1$,"c:\",#PB_Program_Wait| #PB_Program_Hide)
startf = GetTickCount_()
timdec=timdec+(startf-starti)
If ReadFile(0,"c:\"+GetFilePart(nfile$(i))):lufa2=Lof (0):mefile2 =AllocateMemory(lufa2):ReadData(0,mefile2,lufa2):C loseFile(0):EndIf
If CompareMemory(mefile,mefile2,lufa1)=0:MessageReque ster("Information"," TEST FAILED COMP/DEC !", 0):End:EndIf
FreeMemory(mefile)
FreeMemory(mefile2)
If (timenc/1000>1800) MessageRequester("Information"," TEST FAILED INT TIME LIMIT!", 0):End:EndIf
If ReadFile(0,nfile$(i)):lufi=Lof(0):total+lufi:Close File(0):EndIf
If ReadFile(0,"c:\t.zip"):lufi=Lof(0):lufi2+lufi:Clos eFile(0) : DeleteFile("c:\t.zip"):EndIf
Testo$=exe1$+para1$+"name="+nfile$(i)+" "+Str(lufa1)+"->"+Str(lufi)+" "+Chr(13)+Chr(10)
If OpenFile(0,"c:\statistica.txt"):FileSeek(0,Lof(0)) :WriteString(0,Testo$):CloseFile(0):EndIf
If ReadFile(0,"c:\"+GetFilePart(nfile$(i))):CloseFile (0) : DeleteFile("c:\"+GetFilePart(nfile$(i))):EndIf
DrawText(0,20+i*15," nome="+nfile$(i)+" result="+Str(lufi)+" ")
Next

DrawText(0,20,"numero file="+Str(nufile)+" total="+Str(total)+" compr="+Str(lufi2)+" TIME="+Str(timenc)+" ",0)
Testo$="number of files="+Str(nufile)+" original="+Str(total)+" compressed="+Str(lufi2)+" ENC="+StrF(timenc/1000)+" sec. DEC="+StrF(timdec/1000)+" sec."+Chr(13)+Chr(10)
If OpenFile(0,"c:\statistica.txt"):FileSeek(0,Lof(0)) :WriteString(0,Testo$):CloseFile(0):EndIf
MessageRequester("TEST COMPLETED!","num files="+Str(nufile)+" tot="+Str(total)+" C="+Str(lufi2)+" ENC="+Str(timenc)+" DEC="+Str(timdec), 0)
End
; P.S. Place Arc.exe in c:\

[Nania Francesco]

Also try you bulat my program of test and I believe that being automatic nothing is tied up to my intervention!!!! Hi!

[Bulat Ziganshin]

1. arc.groups should be put in the same directory as arc.exe. well, i think that archiver tester should be able at least to read english docs which describes installation procedure

2. this script compress each file into separate archive

overall, i think that MOC is the most meaningless test i ever seen

[Bulat Ziganshin]

Nania, i use ruby script for automatic archiver testing. this script is included in freearc distribution, but i don't think that your test with such script will be better. using nonsolid compression is so obvious mistake that i can't trust your tests

[Nania Francesco]

I am interested to make a will everything, tell me Bulat as can make a will with a solo I command in formed "solid" all the folders of MOC and the files with arc.exe and me I will do him/it if is more powerful person in compression!

[Bulat Ziganshin]

such questions should be asked before making test. if you don't know how to use archivers - don't make a "test"

Nania, your compression algorithms are great, but this test is below any critique

[pat357]

@ Nania
I have to agree with Bulat on this...
Why don't you make a simple test like

timer arc a -mx -r archive .\testfiles\*
and
timer arc t archive.arc

You don't need a script to do this, do you ?

Jest test this and let us know the timings and ratio's.

[Bulat Ziganshin]

the comparison should AUITOMATICALLY test HUNDREDS of programs/modes. one commandline for handful testing of one freearc mode cannot solve problem - tester himself should know how to run various archivers, how to measure speed and so on. Nania's test is just source of confusion - noone knows what he really test

[Nania Francesco]

AT MOMENT FREEARC IS IN TESTING..... [SOLID] GREAT COMPRESSION !

[Nania Francesco]

UPDATED FreeARC 0.50 Solid [option -r ] - m1, m3,m5,mx

[Bulat Ziganshin]

i think that freearc -m2 may be "most efficient compressor" in your rating

but actually i wonder how many compressors was tested in the same fashion, in particular, it seems that all one-file compressors are tested in non-tar mode

[Nania Francesco]

Please write the best options for you in console mode !
Tar active ....?
Detection file level....?
Multimedia compression....?
P.S. More compressor as tested in NON-TAR-MODE!
Tar mode is for RAR - 7ZIP -FREEARC -WINTURLE - SQUEEZ -DCGA

[Bulat Ziganshin]

best modes are -m1..-m5 and -mx

[Nania Francesco]

I want to know if I have to for instance add also other commands

Arc -r - mx -ma2 -mm=on archive sourcefile

or if it is enough only

arc -mx -r archive sourcefile

[Bulat Ziganshin]

P.S. More compressor as tested in NON-TAR-MODE!
Tar mode is for RAR - 7ZIP -FREEARC -WINTURLE - SQUEEZ -DCGA

if you mean "most compressors" - this makes test absolutely useless