This project is dedicated to the memory of William Morris (aka Frags), who was the main contributor to the bounty but was unable to see the final result.

Sunday, February 26, 2012

Test Kickstarts are kickin'

Finally, I managed to publish the fully working build environment for creating Kickstart files, you can find it in the project's SourceForge Subversion repository.

If you are looking for the binaries for these tests then check out the downloads.

Please make sure that if you would like to fool around with these files then you read the post about them here and the README files in the folders.

Some more tests are coming as soon as I have time to clean up the sources.

14 comments:

  1. AGA chipset is required for the mandelbrot test. I noticed you definitely need to use 68020 (not 040/060) for the JIT to work at the moment.

    Also, ppc.use_tbc=false is required, due to changes in sleep.h. The calculation was never correct anyway, and was always better turned off, but it definitely needs to be turned off for these tests.

    I think i'll just have it ignore this configuration setting and default it to false.

    ReplyDelete
    Replies
    1. The processor type has nothing to do with the JIT emulation, I was simply too lazy to create any sophisticated code in the test codes to turn the cache on for other processor types than 68020.
      When the cache is off then the JIT is turned off. So, if you change the cache enabling at the beginning of the test codes to what is needed by the 040/060 processors then it will work with these too.
      Not that it would matter at all, from the point of view of the JIT the processor type makes no difference, except the supported instructions. (Unsupported instructions lead to exception, which is for maintaining compatibility - this is not implemented yet, but the intention is there in the code. ;)

      Regarding the use_tbc config: I don't think this would ever work on AmigaOS (for PPC), reading the time base register directly is not a wise idea. The change in the sleep function is not absolutely required, probably it would be better to use the same thing what is in the libamiga: do a proper call to the timer.device. I will fix this later on, let's file a bug report! :D

      Delete
  2. Replies
    1. Not really. I haven't had much time recently, and that went to chasing the bug what prevents the OS running on the JIT. I still couldn't find the problem, so I guess I go back to implementing the compiling instead. :-P

      Delete
  3. sounds really bad

    ReplyDelete
    Replies
    1. The nay-sayers say nay. :)

      Compiling implementation is progressing, so just be patient.

      Delete
    2. I m not a nay-sayers i just said its bad.

      Delete
  4. Hi Rachy !
    Have any progress with bug-hunting lately ?:)

    ReplyDelete
    Replies
    1. I gave up the bugfixing for now and went back to implementing instructions. I have a little bit of time now, so I hope I can improve the compiling too.

      Delete
    2. Any chance we get a working version this month?

      Delete
    3. A working version of the code compiler that runs test codes: possibly.

      A working version that is able to run the OS: not very likely.

      Delete