Back to benchmark

Thank to following people who helped me to get benchmark where it is:
LovePimple - for his unstoppable optimism :)
Idris "nimdamsk" Kubatayev - for a way to squeeze benchmark tables a lot
Ilia "encode" Muraviev - for encode.ru
Fallon - for design suggestions
Christian Martelock - for design suggestions
Uwe Herklotz - for putting dynamic benchmarking idea into my head

Information about testing computer:
CPUAMD Athlon 64 X2 3800+ dual-core, L1: 64 + 64 kB/core, L2: 512 kB/core (not overclocked)
RAM2GB DDR2-666
OSWindows XP Professional SP1 (32-bit)
HDDSATA Seagate Barracuda 7200.9 (160 GB)

Information about testset
Codename Information (pictures shown are only thumbnails of parts of the testset)
EXE English application. Some BMPs inside.
TXT1 English fiction book.
TXT2 Datafile - small example:
"pufd"="_",<"pu"="_">
"pug"="_",<"pu"="_">
"pugb"="__"
"pugd"="_",<"pu"="_">
"pugd"="_",<"pu"="_">
DLL A dynamic-linked library.
XM A music module.
MP3 An UK Garage style MP3.
PDF A reference manual (English).
PSD File has been released into public here.PSD thumbnail
BMP Screenshot of a Windows desktop full of icons.BMP thumbnail
PNG Screenshot of many folders opened over themselves.PNG thumbnail
PPT A Powerpoint presentation. Czech text, some JPGs and GIFs inside.
SAVE Save from game Gothic III.

Additional information:
Each compressor is tested for maximal compression, not for speed nor efficiency.
Each file is compressed into separate archive and total size is a sum of those archives.
Test files are not available online (with exception of PSD).
Compression and decompression speeds computed using Timer 3.01 by Igor Pavlov (global/wall time).