Results 1 to 5 of 5

Thread: Test set: Java application

  1. #1
    Member m^2's Avatar
    Join Date
    Sep 2008
    Location
    Ślůnsk, PL
    Posts
    1,611
    Thanks
    30
    Thanked 65 Times in 47 Posts

    Test set: Java application

    The test set is Dimes, a java application.
    Almost 3/4 of size is zipped already, so overall compression ratio is poor.
    Code:
    Files       Size        Files no.
    Total       10569844    103
    .jar        7335360     13
    Exacutables 1465270     15
    text        1197959     23
    .ico        410290      7
    .zip        68233       1
    .jpg        42639       5
    .gif        33267       29
    .class      11203       5
    others      5623        5
    The best results:
    Code:
    Archiver                Size    Time
    quick -1                8375483 0.093
    quick -0                8341053 0.125
    quick -2                8296855 0.140
    quick -3                8216127 0.234
    NanoZip -cf             8006016 0.328
    thor e3                 7920016 0.453
    slug                    7899756 0.468
    NanoZip -cd             6991119 0.671
    FreeArc -m2             6941817 0.984
    NanoZip -cD             6895695 1.422
    FreeArc -m3             6888428 1.890
    FreeArc -m3 -ms         6880879 2.687
    FreeArc -m4 -ms         6854155 2.813
    FreeArc -m7x -ms        6850379 3.328
    FreeArc -m9x -ms        6850379 3.328
    FreeArc -m4             6838437 3.390
    FreeArc -m5 -ma- -ms    6834065 4.594
    FreeArc -max -ms        6814810 6.141
    FreeArc -max            6805760 7.719
    CCM 2                   6787225 11.500
    CCM 3                   6781485 11.875
    CCM 4                   6777505 12.250
    CCM 5                   6774772 12.688
    CCMX 5                  6771811 13.781
    FreeArc -max -ma- -ms   5164156 26.782
    As expected, precomp helps a lot...if you let it do the job, it seems to be used by FreeArc only with -ma- switch.
    Attached Files Attached Files

  2. #2
    Programmer schnaader's Avatar
    Join Date
    May 2008
    Location
    Hessen, Germany
    Posts
    551
    Thanks
    206
    Thanked 182 Times in 87 Posts
    Quote Originally Posted by m^2 View Post
    The test set is Dimes, a java application.
    Almost 3/4 of size is zipped already, so overall compression ratio is poor.
    Code:
    Total       10569844    103
    .jar        7335360     13
    [...]
    FreeArc -max -ma- -ms   5164156 26.782
    As expected, precomp helps a lot...if you let it do the job, it seems to be used by FreeArc only with -ma- switch.
    I downloaded the latest installer (0.5.0) and couldn't manage to get the exact same file count, but for example JAR count and size is exactly the same, so I guess it can be compared. My version has 357 files and 10853727 bytes.

    Archiving everything using UHARC and compressing with CCM 2 gives 6928907 bytes.
    Using one Precomp run and CCM 2 gives a result near to your FreeArc result: 5140795 bytes.

    But for JAR archives, two runs should be used. The first extracts JAR content, the second extracts zLib streams in PNGs/GIFs and recompresses JPGs that were inside the JAR. And indeed, using two Precomp runs gives amazing 2542845 bytes!

    The third run detects some further PNGs and GIFs and comes down to 2536746, but that's not really worth it.
    http://schnaader.info
    Damn kids. They're all alike.

  3. #3
    Member m^2's Avatar
    Join Date
    Sep 2008
    Location
    Ślůnsk, PL
    Posts
    1,611
    Thanks
    30
    Thanked 65 Times in 47 Posts
    Damn, I run just 1 iteration with my latest testing, I'll have to redo this.
    News:
    Tornado p1..4
    m1_paq
    kzip
    paq9a
    SBC (more of it)
    PAQ8p

    Precomp+PAQ8P 4309076 bytes.

    After some investigation, I noticed that DIMES keeps some logs in the Program Files.
    I tested on a copy, so all archivers got exactly the same input. I can publish the test set and I'll probably retest everything with a clean install one day.

    ADDED:
    I just tried 3 passes of precomp + paq8p1.
    1704947. Incredible. The best result w/out precomp so far is 6629922 (paq8p -4, I didn't try higher modes yet).
    I'll post updated spreadsheet later, I want to add at least nanozip 0.5 results.
    3 * Precomp + CCM2:
    2368761. Smaller than yours.
    Attached Files Attached Files
    Last edited by m^2; 23rd October 2008 at 16:26.

  4. #4
    Member m^2's Avatar
    Join Date
    Sep 2008
    Location
    Ślůnsk, PL
    Posts
    1,611
    Thanks
    30
    Thanked 65 Times in 47 Posts
    Updated.
    News: Nanozip 0.05
    Qpress 0.22
    3-pass Precomp + PAQ

    Code:
    Archiver                    Size    Time
    quick -1                    8375483 0.093
    quick -0                    8341053 0.125
    quick -2                    8296855 0.140
    quick -3                    8216127 0.234
    NanoZip -cf                 8007565 0.328
    thor e3                     7920016 0.453
    slug                        7899756 0.468
    NanoZip -cd                 6990581 0.656
    FreeArc -m2                 6941817 0.984
    NanoZip -cD                 6894397 1.390
    FreeArc -m3                 6888428 1.890
    FreeArc -m3 -ms             6880879 2.687
    FreeArc -m4 -ms             6854155 2.813
    FreeArc -m9x -ms            6850379 3.328
    FreeArc -m7x -ms            6850379 3.328
    FreeArc -m4                 6838437 3.390
    FreeArc -m5 -ma- -ms        6834065 4.594
    FreeArc -max -ms            66814810 6.141
    FreeArc -max                6805760 7.719
    CCM 2                       6787225 11.500
    CCM 3                       6781485 11.875
    CCM 4                       6777505 12.250
    CCM 5                       6774772 12.688
    CCMX 5                      6771811 13.781
    FreeArc -max -ma- -ms       5164156 26.782
    PRECOMP+PAQ8p1 -7           4309076 2950.328
    PRECOMP(3 passes)+PAQ8p1 -7 1704947 4000.500
    EDIT: Why do I always forget to attach the results?
    Attached Files Attached Files

  5. #5
    Programmer schnaader's Avatar
    Join Date
    May 2008
    Location
    Hessen, Germany
    Posts
    551
    Thanks
    206
    Thanked 182 Times in 87 Posts
    Quote Originally Posted by m^2 View Post
    3 * Precomp + CCM2:
    2368761. Smaller than yours.
    I guess our test sets still differ a bit, furthermore I archived files with UHArc which compresses filenames and this reduces compression ratio a bit.

    By the way, 1.7 MB is really impressive - I still get surprised about how well JAR archives can be recompressed.
    http://schnaader.info
    Damn kids. They're all alike.

Similar Threads

  1. GearEnc - test application gear encoding
    By Sportman in forum Data Compression
    Replies: 7
    Last Post: 19th May 2013, 04:33
  2. An interesting test set
    By nanoflooder in forum Data Compression
    Replies: 12
    Last Post: 13th April 2009, 01:33
  3. Test set: bookstar
    By m^2 in forum Data Compression
    Replies: 5
    Last Post: 11th February 2009, 16:49
  4. Test set: installer
    By m^2 in forum Data Compression
    Replies: 10
    Last Post: 11th February 2009, 13:47
  5. Test set: bitmap
    By m^2 in forum Data Compression
    Replies: 28
    Last Post: 13th January 2009, 16:44

Posting Permissions

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