Find the answer to your Linux question:
Page 1 of 2 1 2 LastLast
Results 1 to 10 of 15
I used PovRay3.5 to benchmark my computer which has a PII450 CPU,G400 MAX DH,192M Memory.The benchmark files are benchmark.pov and benchmark.ini.Here is the result: Windows Xp: 13313 sec Linux 2.4.20: ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #1
    Just Joined!
    Join Date
    Sep 2003
    Posts
    27

    PovRay benchmarks in Linux and Windows


    I used PovRay3.5 to benchmark my computer which has a PII450 CPU,G400 MAX DH,192M Memory.The benchmark files are benchmark.pov and benchmark.ini.Here is the result:
    Windows Xp: 13313 sec
    Linux 2.4.20: 15391 sec
    My XFree86 is 4.3.0,windowsXP and Linux native G400 device driver,and Linux used mga.o with DRI supported .When I login after boot,Linux used 8M memory.I wonder it is the problem that refer to difference between PovRay's linux version and windows,or in their display drivers,or something wrong with my Kernel config?

  2. #2
    Linux Guru
    Join Date
    Oct 2001
    Location
    Täby, Sweden
    Posts
    7,578
    I haven't heard much about Matrox support, actually, but as far as I know, nVidia is about the only 3D accelerator vendor that provides good Linux drivers. So, while I don't know much about matrox, my bet would be that the drivers simply don't cut it.

  3. #3
    Just Joined!
    Join Date
    Sep 2003
    Posts
    27
    I think the PovRay's benchmark mainly tests the floating point computing of a system, because of half a hour big diff, do you think the display driver is a main factor? I wonder whether there is something wrong with my kernel . I use the TWM window manager,and my kernel is built with MTRR support, but no math emulation which seems no use with PII.

    There are Matrix Linux driver in its official web site ,and also ATI has its Linux driver, so I will try it recent day.

  4. #4
    Just Joined!
    Join Date
    Sep 2003
    Posts
    27
    I used Matrox Driver 3.0beta and turned the AGP4x option on,but It seemed useless and the total render time only got down 30 secs.

  5. #5
    Linux Guru
    Join Date
    Oct 2001
    Location
    Täby, Sweden
    Posts
    7,578
    Can't you get info from povray itself on what it is that's taking so long?

  6. #6
    Just Joined!
    Join Date
    Sep 2003
    Posts
    27
    I got no idear of that. I used verbose output in both OS,but both of them seemed the same except the render time. I think I miss something in my kernel which affect floatpoint computing a lot.Because I turned on something like Mtrr...so I don't believe it is linux's fault.

  7. #7
    Linux Guru
    Join Date
    Oct 2001
    Location
    Täby, Sweden
    Posts
    7,578
    The kernel really has nothing with floating-point operation to do, except if you don't have a FPU, in which case it emulates it. If you have a FPU, processes access it with nothing but normal instructions. Therefore, I don't think that that's the answer.

    On the other hand, couldn't it be that something else is running in the background?

  8. #8
    Just Joined!
    Join Date
    Sep 2003
    Posts
    27
    I don't think it's possible.I know clearly about my linux and scripts all of which was written by me,because I built this Linux from scratch. I will test it in a newly installed RH9.0 later to see if it my LFS's fault.

  9. #9
    Just Joined!
    Join Date
    Sep 2003
    Posts
    27
    I used RH9.0 for test,but it was totally useless.I found something said" Peak memory used 5449516bytes" after render.It is means that I have to tune the kernel to improve performace?

  10. #10
    Linux Guru
    Join Date
    Oct 2001
    Location
    Täby, Sweden
    Posts
    7,578
    It's really hard to tell without knowing what part it is that takes so long. Can't you find any way of making it tell you that?
    Is the system swapping at all while you're benchmarking?
    Of course, one thing that you can try is to run the benchmark with /usr/bin/time, which will tell you how much time was spent in user mode vs. kernel mode.

Page 1 of 2 1 2 LastLast

Posting Permissions

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