lzpm 0.15 is here!

[encode]

OK, a new version of the LZPM has been released!

New version introduces pre-built-table based arithmetic encoding. Possibly, this version has the fastest decompression compared to ALL previous versions.

http://lzpm.encode.su/

[LovePimple]

Thanks Ilia!

Mirror: Download

[Nania Francesco]

Thanks Encode!

[Nania Francesco]

LZPM 0.15
SFC Test
Mode 1=13.069.229 in 13,438 sec.
Mode 9=12.535.867 in 438,63 sec.

[Shelwien]

http://shelwien.googlepages.com/lzpm015.htm

Same CeleronM 1.73Ghz.
Also introducing my metric for compressor comparison -
file distribution cputime, calculated like this:

speed = 512000 // 512kbps
downloads = 10
m = ctime+(csize*8/speed+dtime)*downloads

Its cpu computing time spent for encoding the data
once and then 10 times downloading it at 512Kbps and decoding.
Imho quite a realistic metric, as its normally done
this way with most normal compressors, though ideally
compression and decompression time could be overlapped
with transmission time.

[encode]

Thanks for testing!

[Matt Mahoney]

Compression is a little worse than 0.14 but decompression is faster.
http://cs.fit.edu/~mmahoney/compression/text.html#2291

[encode]

Thanks a lot!

By the way, on some files, including text ones, the compression of 0.15 is even better compared to 0.14. Not the case with ENWIKs.

[Nania Francesco]

From Maximum compression:
Simon Berger
Location:
Germany
posted: 02/01/2008 06:28 PM
--------------------------------------------------------------------------------

Sorry for beeing too lame again...

Thanks for testing UhaRc, but there is no decompression result and you didn?t put it in a list.

Is it right that you don?t use a dynamic method for all your pages? - Because then my request should be needless I fear. If you do is it possible to add an additional efficiency with decompression speed because it?s in my opinion for practical use more important then the compression speed. I know that this is a maximum compression page and it is your main interest, but I anyway try to ask :P

Did LZPM 0.15 failed decompression? or didn?t you run it?

Moderator-Comment: I will run the uharc decompression also, but it will not be included in the list as it scores worse then the uharc results already in the list.

All pages are generated dynamic, and yes I will add a decompression time ranking, but I first have to finish all decompression tests (still some backlog).

LZPM: Decomp time 99999 means : archiver failed in decompression !

[encode]

I noticed. I think it's again problem with compatibility with his old computer...
I think I should recompile the LZPM, however, I will do that only after I'll receive an additional information from him...

[Nania Francesco]

I imagined Ilia I think both a problem of compatibility also of System OS! Hi! Good work!

[encode]

Moderator-Comment: Decompression is just fine, I made a major blunder entering the data into the database (did try to automate the process a bit to much). Sorry about that

All OK!