Page 1 of 4 123 ... LastLast
Results 1 to 15 of 46

Thread: 2.5 Ghz G5 Disproportionately Faster in LW 8?

  1. #1

    2.5 Ghz G5 Disproportionately Faster in LW 8?

    Has everyone seen Apple's benchmark of LW8 with the new dual 2.5 Ghz G5s?



    I'm no math wiz, but isn't that speed increase disproportionately faster for the dual 2.5 Ghz G5 over the dual 2 Ghz G5?

    The dual 2 Ghz numbers are pretty much the same numbers you see on Chris' Lightwave benchmarks - the dual 3.2 Ghz Xeon being a bit faster. But the dual 2.5 Ghz G5 blows them both away.

    Am I wrong here? Thanks

  2. #2
    Without thinking about math, a 2x2.5 machine has, essentially, 1.0 GHz processing power more than a 2x2.0 machine, which is basically a 50% increase in power. If the real world performance is about 40% more than the 2x2.0, it would be accurate.
    "One of the things we really have to impress upon people is that it is never good to be fed on by a blood-sucking insect"
    Dr. Anthony Marfin, CDC

  3. #3
    obfuscated SDK hacker Lightwolf's Avatar
    Join Date
    Feb 2003
    Location
    Stuttgart, Germany
    Posts
    13,613
    mlinde, I can't follow you maths here:
    A 2x2.5GHz has a combined 5GHz, which is 1GHz above a dual 2GHz summed up to 4GHz.
    So, as a percentage, you only get 25% more performance.

    Or, skip the dual and just do the maths with a single processor, same thing

    Cheers,
    Mike

  4. #4
    But you don't get a 100% increase in speed with a second processor, usually, so adding the total Ghz and treating that as a single processor is not quite accurate, correct me if I'm wrong.

    In any case, the speed increase of the dual 2.5 Ghz in LW 8 is greater than the Mhz increase, it seems.

  5. #5
    Originally posted by tallscot
    But you don't get a 100% increase in speed with a second processor, usually, so adding the total Ghz and treating that as a single processor is not quite accurate, correct me if I'm wrong.

    In any case, the speed increase of the dual 2.5 Ghz in LW 8 is greater than the Mhz increase, it seems.
    could it be that the Dual 2.5 Ghz faster bus plays a role?

    the new bus is 1.25 ghz/process, to the 1Ghz bus speed per chip on the dual 2.

    J
    _______________________________________
    Discuss sustainability at www.ThinkPlan.org

  6. #6
    if we put this into seconds we might get a clearer comparison - if the baseline was 100 sec., the 2x2.5 would be just under 50 sec., and the 2x2 would be about 62 sec., making the 2x2.5 25% faster than the 2x2.

    Considering that you never really get a proportional mhz / speed increase, this is a good jump in speed ~
    Confirmed -
    No Weapons of Mass Destruction
    or links to Al Queda or 9/11. (Sep. 2003)

  7. #7
    Another bench - not as much of a speed difference
    Attached Images Attached Images
    Confirmed -
    No Weapons of Mass Destruction
    or links to Al Queda or 9/11. (Sep. 2003)

  8. #8
    Gold Member Beamtracer's Avatar
    Join Date
    Feb 2003
    Location
    Citizen of the World
    Posts
    2,226
    Lightwolf is correct with the calculations.

    When software is optimized for the G5 it beats any Intel processor. When software is not optimized for the G5 it goes slower.

    I'd guess that Apple has chosen to display the benchmarks that the G5 does well in.
    Moral of the day:
    Worry about your own morals, not someone else's!

  9. #9
    I posted that to compare the 2 G5's, but note that this benchmark test runs 45 different filters ~ and at double the speed, they really don't need to pick and choose results anyway
    Attached Images Attached Images
    Confirmed -
    No Weapons of Mass Destruction
    or links to Al Queda or 9/11. (Sep. 2003)

  10. #10
    I'm not up to speed on how these things work, and I wouldn't give to much credence to a couple of Benchmarks posted by the manufacture. But I'm guessing that other factors also come into play. I vaguely remember someone talking about the bottlenecks and bus speed.

  11. #11
    PC-killer Ge4-ce's Avatar
    Join Date
    Apr 2003
    Location
    Belgium
    Posts
    675
    I will order myself a G5 2x2.5 on June 28th.. the day Steve jobs hopefully announces the new displays.

    I will test it by hand! I do not thrust benchmarks! I all depends on the scene. A real comparison would be testing all benchmark scenes from the content.

    There are so many differences between raytrace, radiosity, hypervoxels,... all depending on optimisation, multithreading etc.

    Note that it's a LW 8 test. there probably is some G5 optimisation in it.
    __________________________________________________ _



    Koen Delbroek
    D4motion
    www.d4motion.be
    [email protected]

  12. #12
    Originally posted by Lightwolf
    mlinde, I can't follow you maths here:
    A 2x2.5GHz has a combined 5GHz, which is 1GHz above a dual 2GHz summed up to 4GHz.
    So, as a percentage, you only get 25% more performance.

    Or, skip the dual and just do the maths with a single processor, same thing

    Cheers,
    Mike
    Yeah, well if I had a cup of coffee for every time I tried to do math before lunch and failed, I'd never sleep. Yes. 25% increase. Thank you for that correction.
    "One of the things we really have to impress upon people is that it is never good to be fed on by a blood-sucking insect"
    Dr. Anthony Marfin, CDC

  13. #13
    obfuscated SDK hacker Lightwolf's Avatar
    Join Date
    Feb 2003
    Location
    Stuttgart, Germany
    Posts
    13,613
    Originally posted by mlinde
    Yeah, well if I had a cup of coffee for every time I tried to do math before lunch and failed, I'd never sleep. Yes. 25% increase. Thank you for that correction.
    Well, the coffee doesn't always help, I've tried that many times myself

    Cheers,
    Mike

  14. #14
    self flagellating LW monk
    Join Date
    Feb 2003
    Location
    philadelphia
    Posts
    1,708
    Originally posted by Ge4-ce
    Note that it's a LW 8 test. there probably is some G5 optimisation in it.
    the current shipping version of LW (version 8) doesn't have any g5 optimizations. Remember, the official word from on high is that they ship 8 first, then they move it to the new compiler, then they optimize.
    I don't know that I trust Apple's benchmarks, But I'd like to hazard a guess. you don't need higher GigaHertz to have better performance. I haven't been following the size of the processors, but isn't this one the first 90nm g5s? that alone would speed things up a bit, but they could have also tweaked other things like cache size, or placement of the parts. It may be that the 2.5 g5 has a floating point performance increase that outstrips the GigaHertz rating.

  15. #15
    Faster bus, faster processor.

    90nm doesn't make it faster, it just means they can make faster chips (correct me if I'm wrong).

    Someone already compared the new dual 2 Ghz G5 that is based on the 90 nm 970FX processor to the older dual 130 nm 2 Ghz G5 that was based on the 970 processor. The results were identical. This was at www.barefeats.com.

Page 1 of 4 123 ... LastLast

Bookmarks

Posting Permissions

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