Jump to content

Gustavo 'Gus' Carreno

Members
 View Profile  See their activity

  • Content Count

    43

  • Joined

  • Last visited

  • Days Won

    1

Posts posted by Gustavo 'Gus' Carreno

  1. Posted

    Hey my challenge peeps!!

     

    We have finally arrived at the end of the challenge!!
    It's been a fun ride, to say the least.
    The final results have been posted, with the usual minor ooopss, of course.

     

    Like stated before, I'll be accepting entries until I'm no longer in this plane of existence.
    A new results table for those entries will be made, and it will include all of the entries, old and new.

     

    Eventually, I'll also publish a results table for 400 Stations and 8 threads.
    Just need a bit of breathing space to get that done. But it will come.

     

    Let's hope this repository, and the effort of the participants, can stand as an example of the power that still resides in Object Pascal.
    Let's hope we can all shut up the nay sayers by merely pointing at it.

     

    All the best for you all in your future Object Pascal endeavours!!

     

    Cheers,

    Gus

    • Like 1
    • Thanks 1

  4. Posted

    Hey my challenge peeps!!


    We are quickly approaching the deadline for our little challenge.
    While the end date is next Friday, I will post the last results after I run the automation on Sat. of Sun.
    You can still drop PRs after Friday, but they will only be merged after the automation is done on Sat. or Sun.

     

    My intention with this repository is to keep accepting entries for the duration of my life time.
    This is the perfect round up of Object Pascal code to demonstrate that it's not dead and can compete with any of the top languages on SO or TIOBE.

     

    I'm also thinking about adding the 8 thread/400 Stations table for comparison with the Java side of things.
    But I guess that Paweld has made a better attempt here.
    More news on this table later.

     

    Cheers,

    Gus

  6. Posted

    Hey my challenge peeps!!


    Georges suggested that we may give a helping hand to the entries that haven't yet done the change from CRLF to LF only.
    If any of you want to have a stab at it, it would be greatly appreciated!!
    And, does anyone know how to contact Lurendrejer Aksen? His entry has been kinda abandoned and I would like to see it working properly, if possible.
    Many thanks to my wonderful participants in advance!!

     

    Cheers,

    Gus

  7. Posted

    Hey my challenge peeps!!


    We are moving from the input file having CRLF to having only LF as the line terminator.
    Next run, the 27th of April, will be mixed to leave enough time for everyone to make the necessary changes.
    The next run after that, the 4th of May, will not be mixed and will expect every entry to have made changes if necessary.
    The SHA256 for said input file has already been updated on the README.md file.

     

    Cheers,

    Gus

  8. Posted · Edited by Gustavo 'Gus' Carreno

    Hey my challenge peeps!!


    I've been thinking about adding an alternative results table in order to compare to the original Java specs of only 400 stations and 8 threads.
    This means that some of the assumptions for the current state of things would probably be crippling in that case.
    What do you guys think of adding a -4 or --400-stations command line switch to your entries?

     

    Cheers,

    Gus

  10. Posted

    Hey my challenge peeps!!


    I was given owner access to https://github.com/1brc after asking to be included on the https://1brc.dev site.
    Yeah, I know, owner... Still don't understand why, but it's where we're at, LOL!! 😄
    Now I'm in a bit of a pickle: I don't have an icon that will represent Object Pascal.
    There's many for Delphi, and many for Lazarus, but none for Object Pascal in general.
    Can anyone help me on this?

     

    Cheers,

    Gus

  14. Posted

    Hey my challenge peeps!!

     

    I've been busy:

    • The automation is now done for both Linux and Windows
    • We are back to the previous output hash: 4256d19d3e134d79cc6f160d428a1d859ce961167bd01ca528daca8705163910
    • I've altered all the entries that were using an internal version of the rounding functions to use the one from the baseline

    Today is Saturday, and now that I have the automation completed, I'll run the full shebang and will update the results table once they are done.

     

    Cheers,

    Gus

  18. Posted · Edited by Gustavo 'Gus' Carreno

    Hey Y'All,

     

    Looks like there is a simple fix: Not using Double but Currency. 

    function RoundExDouble(x: Currency): Double;
begin
  Result := Ceil(x * 10) / 10;
end;

    Does not fix the issue that Delphi is not consistent with Double across Windows 32,64 and Linux 64.

    But at least we now have consistency on our end!!!

     

    Many thanks to paweld for spotting the fix!!

     

    Cheers,

    Gus

  20. Posted

    Hey Y'All,

    After a bit of talking to a prospect entry person, I've come to the realisation that we had a very complicated rounding implementation.

    The conclusion of that talk and further code is now the new official rounding code.

    I've also altered the README.md file to include the code and the new SHA256 hash of the output, alongside an archived file containing the new baseline output.

    Cheers,
    Gus

  24. Posted

    Hey Y'All,

    We are now into the 2 seconds region: 

        ******** Run All ********
     
    ===== Arnaud Bouchez ======
    -- SSD --
    Benchmark 1: abouchez
      Time (mean ± σ):      2.472 s ±  0.061 s    [User: 27.787 s, System: 1.720 s]
      Range (min … max):    2.386 s …  2.588 s    10 runs
     
    ===========

    That's 1 warmup, 10 runs, 16 threads on SSD

    I'm having issues with Linux watchdogs with 2 entries when run on HDD killing my shell.
    Need to solve that and I'll have results for those.

    Cheers,
    Gus

    • Thanks 1
×