benchmark result - SGI

This is a discussion on benchmark result - SGI ; Hi, I ran a benchmark on my R12000 360Mhz Octane yesterday and it said it was about 361 Vax mips. I was using dhrystone benchmarking C prog. Can't remember where I downloaded it from but I've got it bookmarked on ...

+ Reply to Thread
Results 1 to 6 of 6

Thread: benchmark result

  1. benchmark result

    Hi,
    I ran a benchmark on my R12000 360Mhz Octane yesterday and it said it was
    about 361 Vax mips. I was using dhrystone benchmarking C prog. Can't
    remember where I downloaded it from but I've got it bookmarked on my
    Octane (using the laptop at the mo). Anyway the Octane scored 360mips but
    an HP 9000 B2000 with jobs running on the system (single 400Mhz cpu, 1Gb
    mem PA-8500 cpu) scored over 400! I would have thought that an Octane chip
    would have been faster than an HP 9000. I was thinking of buying an HP
    9000 C3000 workstation for messing about with HP-UX 11 on. Does anyone
    have an opinion on HP 9000's? I've worked with them a bit and have an old
    HP 9000 G50 mainframe at home thats like 6 foot tall and bends teh floor
    boards (had to move it into the shed when I moved outa home summer just
    gone). We have a suit of B2000's (one of the ones I benchmarked) at Oxford
    but I don't get to play with them often. The screen are really blury and
    out of focus. Also the OS is set up in such a way that it makes my cry and
    crindge. Wonder as even the B2000's quite a bit faster than a fast Octane
    how it would do at playing back divx under mplayer? The HP's are a
    fraction of the cost of a half decant SGI. I've seen them go for under
    300! Thats a 400Mhz box with 1Gb memory and LDV scsi drives. One 9Gb
    drive and 1 18Gb LDV scsi drive.

    *********************
    Khalid Schofield
    System Administrator / EM Technician
    Dept. Of Materials
    University Of Oxford
    Parks Road
    Oxford
    OX1 3PH

    Email: khalid.schofield@materials.ox.ac.uk
    Tel: 01865 273785
    Fax: 01865 283333
    Web: http://www-em.materials.ox.ac.uk/peo...eld/index.html


  2. Re: benchmark result

    In article ,
    Khalid Schofield writes:

    > I ran a benchmark on my R12000 360Mhz Octane yesterday and it said it was
    > about 361 Vax mips. I was using dhrystone benchmarking C prog. Can't
    > remember where I downloaded it from but I've got it bookmarked on my
    > Octane (using the laptop at the mo). Anyway the Octane scored 360mips but
    > an HP 9000 B2000 with jobs running on the system (single 400Mhz cpu, 1Gb
    > mem PA-8500 cpu) scored over 400! I would have thought that an Octane chip
    > would have been faster than an HP 9000.


    You are much better off comparing real application performance or at least
    use some application based benchmark instead. But yes, PA Risc machines give
    pretty good bang for the buck compared to other used Unix machines when it
    comes to raw CPU power. If price/cpu-performance is such a large measure for
    you, you could just as well go with x86 instead..

    > Wonder as even the B2000's quite a bit faster than a fast Octane
    > how it would do at playing back divx under mplayer?


    Not sure. HP UX is not even mentioned in the mplayer porting status document at
    all. If it works, it will have exactly the same problems as the Octane.
    The machine wastes arouns 4-5x as much cpu time per frame as some x86 linux box
    (at comparable CPU speed) does using Xv video out or the like, where colorspace
    conversions are done in hardware by the GPU.

    I think this could be done in Hardware on SGIs too, using the OpenGL Imaging
    extension, but someone will have to implement this first (please!). For now,
    the Open GL plugins only help with scaling..

    I'm currently playing with mplayer CVS and I think I got all the relevant
    output plugins (ao_sgi, ao_sdl, vo_x11, vo_gl, vo_gl2) in a somewhat working
    state again. I think i'll finally update my aging 0.90rc3 package when 1.0pre3
    is out..

    so long,
    Timo

    --
    Timo Kanera . GPG Key-ID: 1024D/30CDB412



  3. Re: benchmark result

    Yeh I'm waiting for the next ver of mplayer. It's getting better all of
    the time for us irix users but the apps still arn't using the hardware
    that well at all :-( I think I'll bight the bullett and go get a PeeCee
    from Mr Dell for use at work some time soon. I'm testing passwords using
    john the ripper so my desktop isn't that great at it. Just to mention that
    I'm the IT officer in the electron microscopy group at oxford uni. I gave
    my PC to a user and decided I could cope just using a 4 cpu dec alpha
    server 2100. Used to hold 12000 peoples email accounts on it from 1996 to
    2001. Not it's running FreeBSD in my office. Not very fast for number
    crunching though. I'd love to get funding to build an overclocked P4 and
    ram it to 5Ghz using the vapo-chill case that has a refridgeration plant
    in it. That'd do the job for a few months until a faster chip came out.

    I havn't managed to get mplayer working with the -vo gl2 option yet :-(
    What do I need to get it working? What packages and devices do I need?
    I've got the mplayer -vo gl working fine.



    *********************
    Khalid Schofield
    System Administrator / EM Technician
    Dept. Of Materials
    University Of Oxford
    Parks Road
    Oxford
    OX1 3PH

    Email: khalid.schofield@materials.ox.ac.uk
    Tel: 01865 273785
    Fax: 01865 283333
    Web: http://www-em.materials.ox.ac.uk/peo...eld/index.html


    On Sun, 9 Nov 2003, Timo Kanera wrote:

    > In article ,
    > Khalid Schofield writes:
    >
    > > I ran a benchmark on my R12000 360Mhz Octane yesterday and it said it was
    > > about 361 Vax mips. I was using dhrystone benchmarking C prog. Can't
    > > remember where I downloaded it from but I've got it bookmarked on my
    > > Octane (using the laptop at the mo). Anyway the Octane scored 360mips but
    > > an HP 9000 B2000 with jobs running on the system (single 400Mhz cpu, 1Gb
    > > mem PA-8500 cpu) scored over 400! I would have thought that an Octane chip
    > > would have been faster than an HP 9000.

    >
    > You are much better off comparing real application performance or at least
    > use some application based benchmark instead. But yes, PA Risc machines give
    > pretty good bang for the buck compared to other used Unix machines when it
    > comes to raw CPU power. If price/cpu-performance is such a large measure for
    > you, you could just as well go with x86 instead..
    >
    > > Wonder as even the B2000's quite a bit faster than a fast Octane
    > > how it would do at playing back divx under mplayer?

    >
    > Not sure. HP UX is not even mentioned in the mplayer porting status document at
    > all. If it works, it will have exactly the same problems as the Octane.
    > The machine wastes arouns 4-5x as much cpu time per frame as some x86 linux box
    > (at comparable CPU speed) does using Xv video out or the like, where colorspace
    > conversions are done in hardware by the GPU.
    >
    > I think this could be done in Hardware on SGIs too, using the OpenGL Imaging
    > extension, but someone will have to implement this first (please!). For now,
    > the Open GL plugins only help with scaling..
    >
    > I'm currently playing with mplayer CVS and I think I got all the relevant
    > output plugins (ao_sgi, ao_sdl, vo_x11, vo_gl, vo_gl2) in a somewhat working
    > state again. I think i'll finally update my aging 0.90rc3 package when 1.0pre3
    > is out..
    >
    > so long,
    > Timo
    >
    > --
    > Timo Kanera . GPG Key-ID: 1024D/30CDB412
    >
    >
    >


  4. Re: benchmark result

    In article ,
    Khalid Schofield wrote:
    >Hi,
    >I ran a benchmark on my R12000 360Mhz Octane yesterday and it said it was
    >about 361 Vax mips. I was using dhrystone benchmarking C prog. Can't
    >remember where I downloaded it from but I've got it bookmarked on my
    >Octane (using the laptop at the mo). Anyway the Octane scored 360mips but
    >an HP 9000 B2000 with jobs running on the system (single 400Mhz cpu, 1Gb
    >mem PA-8500 cpu) scored over 400! I would have thought that an Octane chip


    A) that's not a significant difference.
    B) making a buying decision based on dhrystone is a very bad idea.

    Circa 1985 dhrystone was pretty much all there was (ignoring
    very expensive proprietary suites), and it was
    not representative of anything much even then.

    I have no idea which system is 'faster', but I can at least
    hope the Octane is faster :-) because your test proves little
    or nothing... Maybe it suggests what compiler options you
    chose :-)


    Regards,
    David B. Anderson http://reality.sgiweb.org/davea

  5. Re: benchmark result

    I have a few HP machines - they are very good systems, but it depends more on
    your application, more so than SGI/IRIX computers. I find SGI/IRIX to be a
    much more "fun" desktop system to play with at home than HP-UX. For a real
    application, this might be different (but Im sure you know this already).
    After getting my octane, I stopped using all my other SUN and HP-UX systems
    for desktop use. To their defense, the octane is much faster than the older
    Sparcs and HP-UX boxes I have, but I like the "feel" that IRIX has much more.

    I have much respect for SUN and HP, but using IRIX comes more as a personnel
    preference I guess.

    BTW. - Just wondering, how much of a real world performace boost are you
    getting with the R12k 360mhz CPU? Have you tested a truly CPU-intensive
    program yet? i.e. xmame!
    http://x.mame.net

    cheers!
    mike

    > I ran a benchmark on my R12000 360Mhz Octane yesterday and it said it was
    > about 361 Vax mips. I was using dhrystone benchmarking C prog. Can't
    > remember where I downloaded it from but I've got it bookmarked on my
    > Octane (using the laptop at the mo). Anyway the Octane scored 360mips but
    > an HP 9000 B2000 with jobs running on the system (single 400Mhz cpu, 1Gb
    > mem PA-8500 cpu) scored over 400! I would have thought that an Octane chip
    > would have been faster than an HP 9000. I was thinking of buying an HP
    > 9000 C3000 workstation for messing about with HP-UX 11 on. Does anyone
    > have an opinion on HP 9000's? I've worked with them a bit and have an old
    > HP 9000 G50 mainframe at home thats like 6 foot tall and bends teh floor
    > boards (had to move it into the shed when I moved outa home summer just
    > gone). We have a suit of B2000's (one of the ones I benchmarked) at Oxford
    > but I don't get to play with them often. The screen are really blury and
    > out of focus. Also the OS is set up in such a way that it makes my cry and
    > crindge. Wonder as even the B2000's quite a bit faster than a fast Octane
    > how it would do at playing back divx under mplayer? The HP's are a
    > fraction of the cost of a half decant SGI. I've seen them go for under
    > 300! Thats a 400Mhz box with 1Gb memory and LDV scsi drives. One 9Gb
    > drive and 1 18Gb LDV scsi drive.



  6. Re: benchmark result

    I do quite a bit of cpu intensive stuff. I play with semper. It's an
    interesting prog wrote in fortran 77 for image work. We use it mainly for
    looking at data produced from our electron microscopes. I backup my stuff
    using gzip and cpio which hammers the cpu when i back up my movie
    collection (460Gb's). I'll have a play with xmame. I downloaded it but
    didn't get round to looking up how to use it to do a benchmark on my box
    and get 'numbers' out of it. I play back divx all the time which needs a
    half decent cpu.

    *********************
    Khalid Schofield
    System Administrator / EM Technician
    Dept. Of Materials
    University Of Oxford
    Parks Road
    Oxford
    OX1 3PH

    Email: khalid.schofield@materials.ox.ac.uk
    Tel: 01865 273785
    Fax: 01865 283333
    Web: http://www-em.materials.ox.ac.uk/peo...eld/index.html


    On Sun, 9 Nov 2003, mike wrote:

    > I have a few HP machines - they are very good systems, but it depends more on
    > your application, more so than SGI/IRIX computers. I find SGI/IRIX to bea
    > much more "fun" desktop system to play with at home than HP-UX. For a real
    > application, this might be different (but Im sure you know this already).
    > After getting my octane, I stopped using all my other SUN and HP-UX systems
    > for desktop use. To their defense, the octane is much faster than the older
    > Sparcs and HP-UX boxes I have, but I like the "feel" that IRIX has much more.
    >
    > I have much respect for SUN and HP, but using IRIX comes more as a personnel
    > preference I guess.
    >
    > BTW. - Just wondering, how much of a real world performace boost are you
    > getting with the R12k 360mhz CPU? Have you tested a truly CPU-intensive
    > program yet? i.e. xmame!
    > http://x.mame.net
    >
    > cheers!
    > mike
    >
    > > I ran a benchmark on my R12000 360Mhz Octane yesterday and it said it was
    > > about 361 Vax mips. I was using dhrystone benchmarking C prog. Can't
    > > remember where I downloaded it from but I've got it bookmarked on my
    > > Octane (using the laptop at the mo). Anyway the Octane scored 360mips but
    > > an HP 9000 B2000 with jobs running on the system (single 400Mhz cpu, 1Gb
    > > mem PA-8500 cpu) scored over 400! I would have thought that an Octane chip
    > > would have been faster than an HP 9000. I was thinking of buying an HP
    > > 9000 C3000 workstation for messing about with HP-UX 11 on. Does anyone
    > > have an opinion on HP 9000's? I've worked with them a bit and have an old
    > > HP 9000 G50 mainframe at home thats like 6 foot tall and bends teh floor
    > > boards (had to move it into the shed when I moved outa home summer just
    > > gone). We have a suit of B2000's (one of the ones I benchmarked) at Oxford
    > > but I don't get to play with them often. The screen are really blury and
    > > out of focus. Also the OS is set up in such a way that it makes my cry and
    > > crindge. Wonder as even the B2000's quite a bit faster than a fast Octane
    > > how it would do at playing back divx under mplayer? The HP's are a
    > > fraction of the cost of a half decant SGI. I've seen them go for under
    > > 300! Thats a 400Mhz box with 1Gb memory and LDV scsi drives. One 9Gb
    > > drive and 1 18Gb LDV scsi drive.

    >
    >


+ Reply to Thread