Results 1 to 13 of 13

Thread: TC 5.0dev5 released! [!]

  1. #1
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    What's new:
    Now TC uses my own, unique, simple and pretty fast order-2-0 PPM. In addition, due to a new PPM, TC have improved compression, and on text files, compression was improved significantly.

    Enjoy!

    Link:
    Download TC 5.0dev5 (26 KB)


  2. #2
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    TC 5.0dev5

    Performance on 'Large Text Compression Benchmark':

    <u>ENWIK8</u>
    compressed size: 29,826,603 bytes

    <u>ENWIK9</u>
    compressed size: 259,748,113 bytes
    compression time: 270 sec.
    decompression time: 267 sec.

    LZP hashtable: 4 MB
    PPM hashtable: 4 MB
    search buffer: 16 MB

    memory usage: 24 MB

    P4 3.0 GHz, 1 GB RAM, Windows XP SP2


  3. #3
    Guest
    Thanks!

  4. #4
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    I'm very curious about the 'Maximum Compression' results!


  5. #5
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    Performance on SFC:

    world95.txt: 671,284 bytes
    fp.log: 684,512 bytes [!]
    english.dic: 853,470 bytes

    acrord32.exe: 1,741,005 bytes
    mso97.dll: 2,097,261 bytes

    rafale.bmp: 1,109,537 bytes
    A10.jpg: 859,787 bytes

    vcfiu.hlp: 748,028 bytes
    ohs.doc: 880,133 bytes
    FlashMX.pdf: 3,826,258 bytes

    total: 13,471,275 bytes (place #81)


  6. #6
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    Working on TC...








  7. #7
    Guest
    Performance on SFC:

    world95.txt: 671,284 bytes
    fp.log: 684,512 bytes [!]
    english.dic: 853,470 bytes

    acrord32.exe: 1,741,005 bytes
    mso97.dll: 2,097,261 bytes

    rafale.bmp: 1,109,537 bytes
    A10.jpg: 859,787 bytes

    vcfiu.hlp: 748,028 bytes
    ohs.doc: 880,133 bytes
    FlashMX.pdf: 3,826,258 bytes

    total: 13,471,275 bytes (place #81)



    Are you happy with these results?

  8. #8
    Guest
    TC 5.0dev5 'Maximum Compression' results will be added to the site this weekend.

    Werner

  9. #9
    Guest
    tc compresses itself to 28K.
    50% original size!

    luis

  10. #10
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    Are you happy with these results?
    Well, this program is designed for speed. In addition, I'm very happy with results on 'fp.log' - 684,512 bytes. Note, 7-Zip 4.42, using 'Ultra' preset compresses this file to 838,823 bytes. Also note, this is not the final version of TC! I have a lots of ideas about improving current Fast PPM: update exclusion, better escape mechanism, etc.

    TC 5.0dev5 'Maximum Compression' results will be added to the site this weekend.
    Thanks!

    tc compresses itself to 28K.
    It's poorer compared to PKZIP, since TC is not so effective on small files. But the difference in compression is not too high though.


  11. #11
    Guest
    188 sec in MFC?

  12. #12
    Guest
    MFC Compression is not as good?

  13. #13
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    188 sec in MFC?

    MFC Compression is not as good?


    Basically, such performance on MFC is not surprise for me.

    Speed
    The speed is significantly affected due to a hashtable. Furthermore, on mixed data this table works significantly slower than on text data. But I can improve speed, minimizing linear search.

    Compression
    In my Fast PPM algorithm I completely eliminate any escape codes. And for selecting coding order I use special heuristic. On text files this trick works well, but on mixed data this stuff wont work - as a result poor compression. Anyway, I improve that...

    In fact, this Fast PPM is not so fast...


Posting Permissions

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