lzpm 0.05 is here!

[encode]

Okay, check out a new version of the lzpm.

What's new:
+ An improved parsing scheme
+ An improved circular hash chains
+ Fully compatible with previous version (decoder unchanged)

Enjoy!

Link:
lzpm005.zip (26 KB)

Note that at some point this is the last version of lzpm, or new version will be available with some real huge gap...

[Nania Francesco]

LZPM v0.05
Pentium IV 3GHZ HT

A10.jpg > 832,282
AcroRd32.exe > 1,630,322
english.dic > 1,031,885
FlashMX.pdf > 3,766,309
FP.LOG > 797,802
MSO97.DLL > 2,010,427
ohs.doc > 836,570
rafale.bmp > 1,094,750
vcfiu.hlp > 715,795
world95.txt > 603,251

Total = 13,319,393 bytes in 35,547 sec.

Enwik8 ->28.895.366

[LovePimple]

Thanks Ilia!

Mirror: Download

Note that at some point this is the last version of lzpm, or new version will be available with some real huge gap...

Are you starting a new project?

[encode]

Are you starting a new project?

I dont. Just will make some break and will keep myself away from data compression due to some business. Note that I planned LZPM as an experimental polygon for my variant of ROLZ compression testing, maybe I need year or two to finish it...

[Vacon]

Hello everyone,

Just will make some break and will keep myself away from data compression

...maybe I need year or two to finish it...

Are you going to tell us, that you will let us stand in the rain???
Probably I combined the two quotings false!?!
What about your engagement in this forum here? Paused as well?


Best regards!

[encode]

Well, I'm staying online. Read mail and answer forum questions is one thing, inventing a new compression and working on programs is another. I just need some new ideas for my LZPM. Making a break will lead to some wanted ideas and more fresh point of view about my software.

[Black_Fox1]

Thanks for the new LZPM, encode!

[Matt Mahoney]

lzpm 0.05 decompression failed on enwik8. Output was the wrong size, 100,000,105 bytes, and differs starting at offset 04744D06 hex.

The compressed file has size 28,895,366 and md5sum 5288bc7620f6dba04a781ae54b185fed

Output file has md5sum cd002534b9f42c804c0fa877d3e1007d
(should be a1fa5ffddb56f4953e226637dabbb36a )

[encode]

I'll fix it ASAP!

WIth 0.05 I used a really risky way to improve hash chaining, with my brief tests all works fine. SO, I know what's the problem....