Page 1 of 10 123 ... LastLast
Results 1 to 30 of 294

Thread: Squeeze Chart 2007

  1. #1
    Member
    Join Date
    May 2008
    Location
    France
    Posts
    78
    Thanks
    436
    Thanked 22 Times in 17 Posts
    Stephan, I start a new topic where you can notify update of your bench.
    It 'll be easier to follow.
    I also have a request : could you make the test's date more visible on your page ? When you open main benchmark page, you see first revision (REV 6.0 - template ?)... you must search the date on the right end of the page (in small character)

  2. #2
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    Yes, you're right. I will put the date above Revision.
    I have added 2 new photographs on hall of fame page:
    Marcel Lemke (WinACE) and Nania (Hook).
    Currently, PAQ8l is being tested. Hook 0.8c and Quad 1.10
    as well as CCM 1.20a are already tested. Those results will
    be online this afternoon (but PAQ8l only partly).
    Thanks for this topic, Mike

    Yours, Stephan

  3. #3
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    Hey Guys,

    the promised main chart is now online

    Yours,
    Stephan

  4. #4
    Member
    Join Date
    May 2008
    Location
    France
    Posts
    78
    Thanks
    436
    Thanked 22 Times in 17 Posts
    Thanks

  5. #5
    Programmer
    Join Date
    Feb 2007
    Location
    Germany
    Posts
    420
    Thanks
    28
    Thanked 150 Times in 18 Posts
    Thank you very much, Stephan!

  6. #6
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,954
    Thanks
    359
    Thanked 332 Times in 131 Posts
    Thank you Stephan!

  7. #7
    Tester
    Nania Francesco's Avatar
    Join Date
    May 2008
    Location
    Italy
    Posts
    1,565
    Thanks
    220
    Thanked 146 Times in 83 Posts
    Thank you Stephan! your benchmark is perfect!

  8. #8
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    Dear readers,

    I just wanted to let you know that I have finished testing PAQ8L;
    atm Hook 0.8e is running and then Quad. I will set the new results
    online this evening. I woke up later today because I watched the
    fight Maske vs Hill last night..


    Cheers

  9. #9
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,954
    Thanks
    359
    Thanked 332 Times in 131 Posts
    Boxen: Henry Maske vs. Virgil Hill - Die Revanche

    Thanks for info!

  10. #10
    Tester
    Nania Francesco's Avatar
    Join Date
    May 2008
    Location
    Italy
    Posts
    1,565
    Thanks
    220
    Thanked 146 Times in 83 Posts
    We wait for all with eagerness! we wait for all this evening!

  11. #11
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    And not to forget: The brand new M99 from Michael Mansicalco,
    that includes MSufsort 3.1beta, arithmetic coding.. from this algorithm the new M03 will be made and it is rumored that M03
    will outperform PPMonstr..

    More this evenig

  12. #12
    Programmer
    Join Date
    Feb 2007
    Location
    Germany
    Posts
    420
    Thanks
    28
    Thanked 150 Times in 18 Posts
    Quote Originally Posted by Stephan Busch
    And not to forget: The brand new M99 from Michael Mansicalco,
    that includes MSufsort 3.1beta, arithmetic coding.. from this algorithm the new M03 will be made and it is rumored that M03
    will outperform PPMonstr..
    Dont tease us like this. M99 is totally new to me. Cant wait to see first results.

  13. #13
    Programmer michael maniscalco's Avatar
    Join Date
    Apr 2007
    Location
    Boston, Massachusetts, USA
    Posts
    109
    Thanks
    7
    Thanked 80 Times in 25 Posts
    Quote Originally Posted by Christian
    Dont tease us like this. M99 is totally new to me. Cant wait to see first results.
    Actually, I wrote M99 back in 1999. I had not bothered to write a full implementation with BWT because I got destracted with MSufSort to create the BWT and then I invented M03 and had not found time.

    I decided to implement it fully before moving on the M03, which is the current project. I *expect* M03 to perform in the same league as PPMonstr in most cases. I never promised that it would outperform it.

    There is lots of room for improving M99 but I am focusing on M03 instead because I know that its much stronger. So its not likely I will find the time to go back and tune M99 for compression or speed.

    Currently, its almost as strong as Dark for compression and, in fast mode its as fast as bzip2 with better compression at the same blocksizes.

    But M03 is context based compression of the BWT and will be far stronger for compression than M99 can achieve. It wont likely be as fast though.

  14. #14
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    I am testing M99 right now, and it's far better than M99 from 2001,
    but unfortunately, I can't use switch combination -m 32m -a;
    M99 tells me then that it could not find source file. Without '-a'
    option everything runs just fine.

  15. #15
    Programmer
    Join Date
    Feb 2007
    Location
    Germany
    Posts
    420
    Thanks
    28
    Thanked 150 Times in 18 Posts
    Hi Michael!
    Quote Originally Posted by Michael Maniscalco
    Actually, I wrote M99 back in 1999.
    Please excuse my mistake. I am interested in compression but Ive somehow lost the overview (because there are so many compressors).
    Quote Originally Posted by Michael Maniscalco
    But M03 is context based compression of the BWT and will be far stronger for compression than M99 can achieve. It wont likely be as fast though.
    Thank you for this explanation. Your context-based BWT algorithm seems to be very powerful. Im really looking forward to your next release of M03.

    Best regards,

    Christian

  16. #16
    Tester
    Nania Francesco's Avatar
    Join Date
    May 2008
    Location
    Italy
    Posts
    1,565
    Thanks
    220
    Thanked 146 Times in 83 Posts
    Hi Michael! Very good work whit M99 ! Faster Compressor!

  17. #17
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    The new benchmark is online now

  18. #18
    Member Fallon's Avatar
    Join Date
    May 2008
    Location
    Europe - The Netherlands
    Posts
    154
    Thanks
    14
    Thanked 10 Times in 5 Posts
    Quote Originally Posted by Stephan Busch
    I have added 2 new photographs on hall of fame page
    Your hall of fame, Stephan, is a good effort to give a face to history.
    And ... some inappropriate comments about bad hairdays from Darwinian- to our present times rise to my lips.

    http://www.michael-maniscalco.com/compression.htm
    Will have a look at the new version. Thanks.
    (for bodily compression, I prefer the Heckler Koch MSG-90A1 to the Barrett M99. The Russians may like their SV-98 best.).

    Quote Originally Posted by encode
    Boxen: Henry Maske vs. Virgil Hill - Die Revanche
    Maske won on points and has retired again right away. With a new bank account.

  19. #19
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts

    and for mind compression the most widely used tools are alcohol and cigarettes

    Yes, but Maske is already rich: He has some Mc Donalds stores which he supervises and he also has some car-selling houses..

  20. #20
    Programmer
    Join Date
    Feb 2007
    Location
    Germany
    Posts
    420
    Thanks
    28
    Thanked 150 Times in 18 Posts
    Quote Originally Posted by Stephan Busch
    The new benchmark is online now
    Thank you very much, Stephan!
    Your benchmark is awesome and very informative! But I think Im repeating myself.

  21. #21
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    Thanks for the flowers

    But also CCM is awesome and it is my favorite at present

  22. #22
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,954
    Thanks
    359
    Thanked 332 Times in 131 Posts
    Strange thing - theoretically, PIM 1.50 must has the same compression as PIM 1.40BETA - bit to bit identical. At your test the results are slightly different! A bug? Anyone can confirm that PIM 1.50 has different compression than 1.40?

  23. #23
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,954
    Thanks
    359
    Thanked 332 Times in 131 Posts
    I've made a few tests - compression of 1.50 is the SAME as with 1.40BETA! I compressed many files, checked compressed size, md5 checksums of archives - all okay! Probably you changed file names in your test?

  24. #24
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,954
    Thanks
    359
    Thanked 332 Times in 131 Posts
    And one more important thing:

    PIM 1.50 is NOT beta!
    QUAD 1.10 is NOT beta!

    At your test we can see:
    PIM 1.50 Beta (2007)
    QUAD 1.10 Beta (2007)



    Please fix that!

  25. #25
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    Hi Ilia,

    I am almost sure I used Pim 1.50 for the test and regarding the testsets: They are always identical, on the same testsets were Pim
    1.40 running.. Indeed strange.. I will check this again and fix the writing errors. Please excuse me, Ilia..

  26. #26
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,954
    Thanks
    359
    Thanked 332 Times in 131 Posts
    Quote Originally Posted by Stephan Busch
    I will check this again and fix the writing errors.
    Thank you very much Stephan!

  27. #27
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    I have retested PIM 1.50 on all testsets and got the same results.
    So the archives are not the same compared to Pim 1.40 beta.
    But I can remove 1.40 results if you want me to do..

  28. #28
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,954
    Thanks
    359
    Thanked 332 Times in 131 Posts
    Quote Originally Posted by Stephan Busch
    I have retested PIM 1.50 on all testsets and got the same results.
    So the archives are not the same compared to Pim 1.40 beta.
    Probably, something wrong with PIM 1.40BETA...

    Quote Originally Posted by Stephan Busch
    But I can remove 1.40 results if you want me to do..
    Okay!

    Anyway, as announced, PIM 1.50 is some kind of final version (however, the popularity of PIM is growing up with a logarithmic progression, so, it is possible some day Ill back to its development).

    For me now QUAD is more important. QUAD 1.11 has a slightly higher compression, however, I stuck in the middle with Optimal Parsing... But the main thing Ive done in this version is the serious compression speed improvements in both modes. Thus QUAD with new Normal mode has an essentially new place, being faster than Fast LZMA and at the same time with Higher compression...

  29. #29
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    cool, I will test Quad 1.11 asap

  30. #30
    Tester
    Stephan Busch's Avatar
    Join Date
    May 2008
    Location
    Bremen, Germany
    Posts
    872
    Thanks
    457
    Thanked 175 Times in 85 Posts
    The fixed benchmark is online. Thanks again for reporting, Ilia

Page 1 of 10 123 ... LastLast

Similar Threads

  1. Squeeze Chart
    By Stephan Busch in forum Data Compression
    Replies: 501
    Last Post: 9th November 2018, 06:05
  2. Squeeze Chart 2006 - 02 April/13 May
    By encode in forum Forum Archive
    Replies: 9
    Last Post: 17th July 2006, 04:39

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •