Results 1 to 11 of 11

Thread: Some Questions regarding Hutterprize

  1. #1
    Member
    Join Date
    Jun 2008
    Location
    G
    Posts
    377
    Thanks
    26
    Thanked 23 Times in 16 Posts

    Some Questions regarding Hutterprize

    Hi,

    I have some questions regarding the Hutterprize:

    1. Is it allowed the have more than one decompressor?
    2. Can I use for a 3rd party decompressor like for example "unzpaq" or "paq8px" plus my own which calls the 3rd party decompressor?
    3. Is it allowed to use a jar file as decompressor? Because .net exe are also allowed?

    Thank you in advance.
    Last edited by thometal; 5th July 2018 at 21:50.

  2. #2
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Switzerland
    Posts
    553
    Thanks
    356
    Thanked 355 Times in 192 Posts
    First of all, sorry for the delay.
    I think noone knows how to answer your question(s) - including myself.
    Do you mean you would like to preprocess enwik8 with a .net or java application then compress the result by some existing compression app?
    Last edited by Gotty; 8th July 2018 at 14:53.

  3. #3
    Member
    Join Date
    Jun 2009
    Location
    Kraków, Poland
    Posts
    1,497
    Thanks
    26
    Thanked 132 Times in 102 Posts
    I think Hutter Prize rules are flawed as they heavily favor programs written in C and C++. Why? Because standard libraries used in C or C++ programs are typically installed in a stock OS install. That's not true for other languages. Therefore if I for example write a program in Rust then I have to link Rust's standard library statically, which adds a lot to executable weight. Dynamic linking against Rust's standard library results in small executables with size comparable to programs written in C or C++.

    Matt offers help with running a C/ C++ program on his computer. If C/ C++ programs get such special treatment then why not apply it to other languages? I propose following rules:
    - standard library is not counted into executable size. Standard library is e.g.: Rust's standard library, Java Runtime Environment, .NET Framework, etc This should apply to popular frameworks with easy installation process so everyone could replicate the test
    - contestant sends two version of his program: one self contained (i.e. statically linked against standard libraries, containing necessary VMs, interpreters, etc) and one that requires installation of standard library (defined in previous point) in the operating system
    - self contained version is tested first and if it doesn't work out of the box it's rejected
    - if the self contained version works as advertised then Matt downloads standard library and tests the lightweight version of decompressor executable

    What do you think?

  4. #4
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Switzerland
    Posts
    553
    Thanks
    356
    Thanked 355 Times in 192 Posts
    Quote Originally Posted by Piotr Tarsa View Post
    - standard library is not counted into executable size. Standard library is e.g.: Rust's standard library, Java Runtime Environment, .NET Framework, etc This should apply to popular frameworks with easy installation process so everyone could replicate the test
    I agree.
    However static linking is not really applicable to java or most interpreted languages let alone script languages. I'd say if the dynamically linked version "runs" then it "passes". If it needs (complains about) a dll not present in the OS, not included in a runtime environment (java, .net) or not included as a dll in a compiler's runtime package (msvc), then it won't pass.

    Thometal, let's say anything is allowed. Can you make a winning entry? If not yet, you have a lot of work ahead of you. Winning the price requires a lot. After that work is done, let's come back to your questions.

  5. #5
    Member
    Join Date
    Jun 2009
    Location
    Kraków, Poland
    Posts
    1,497
    Thanks
    26
    Thanked 132 Times in 102 Posts
    Actually you can compile Java programs to standalone Linux executable, but it will be relatively big: https://www.graalvm.org/docs/getting...#native-images
    Java .class files used to produce native image would be much smaller.

    Can you make a winning entry? If not yet, you have a lot of work ahead of you. Winning the price requires a lot. After that work is done, let's come back to your questions.
    I'm working on it. Project is here: https://github.com/tarsa/demixer but I've stopped publishing my work some months ago, because if I publish them then someone can steal some ideas from it and submit Hutter prize entry before me.

  6. #6
    Member
    Join Date
    Jun 2015
    Location
    Switzerland
    Posts
    876
    Thanks
    242
    Thanked 324 Times in 197 Posts
    Quote Originally Posted by Piotr Tarsa View Post
    I'm working on it. Project is here: https://github.com/tarsa/demixer
    Wonderful news!

  7. Thanks:

    Piotr Tarsa (9th July 2018)

  8. #7
    Member
    Join Date
    Jun 2009
    Location
    Kraków, Poland
    Posts
    1,497
    Thanks
    26
    Thanked 132 Times in 102 Posts
    I wonder how Hutter Prize test machine performance relates to modern CPUs performance. Test machine is Core i7 620M, which is pretty old and new CPU are much faster.

    Example - compressing enwik8 using PAQ8L:
    1 second on Core i5 4670 @ 3.8 GHz = X seconds on Core i7 620M
    (where = means equal amount of processed file)

    What's X?

  9. #8
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Switzerland
    Posts
    553
    Thanks
    356
    Thanked 355 Times in 192 Posts
    https://www.cpubenchmark.net/cpu.php....40GHz&id=1933
    Single Thread Rating: 2154


    https://www.cpubenchmark.net/cpu.php...2.67GHz&id=849
    Single Thread Rating: 1212

    Probably: X≈1.7

    Edit: is the i5 overclocked (3.4 -> 3.8)? If it is, X could be more.
    Edit: for cpu bound compressors it is X, but memory bound compressors would probably depend more on RAM speed.

  10. Thanks:

    Piotr Tarsa (25th July 2018)

  11. #9
    Member
    Join Date
    Aug 2008
    Location
    Planet Earth
    Posts
    1,026
    Thanks
    103
    Thanked 410 Times in 285 Posts
    My test system i7 8700K at 5GHz:

    Single Thread Rating: 3037

  12. #10
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Switzerland
    Posts
    553
    Thanks
    356
    Thanked 355 Times in 192 Posts
    Quote Originally Posted by Sportman View Post
    My test system i7 8700K at 5GHz:
    Single Thread Rating: 3037

  13. #11
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Switzerland
    Posts
    553
    Thanks
    356
    Thanked 355 Times in 192 Posts

Similar Threads

  1. questions about data correction
    By just a worm in forum The Off-Topic Lounge
    Replies: 4
    Last Post: 2nd February 2014, 18:39
  2. FreeArc usability questions
    By TheEmptyMind in forum Data Compression
    Replies: 12
    Last Post: 14th July 2013, 14:36
  3. Greetings, Questions, and Benchmarks
    By musicdemon in forum Data Compression
    Replies: 4
    Last Post: 8th January 2012, 21:45
  4. Questions about compression
    By 0011110100101001 in forum Data Compression
    Replies: 12
    Last Post: 8th December 2011, 01:31
  5. Bunch of stupid questions
    By chornobyl in forum Data Compression
    Replies: 28
    Last Post: 6th December 2008, 17:26

Posting Permissions

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