Intel vs G5 with Universal Binary applications

Discussion in 'PowerPC Macs' started by mrfaize, Nov 28, 2007.

  1. mrfaize macrumors member

    Joined:
    Sep 16, 2007
    #1
    Do Universal Binary applications (Adobe Illustrator CS3) run quicker on Intel than on Power PC if the machines were similary spec'd maybe due to better architecture etc. E.g. a "G5 PPC 2.0Ghz" vs "iMac C2D 2.0ghz"

    I'm looking for real world application use rather than benchmarks.

    wondering if it is better to go second user on a G5 or splash out on a new iMac/MacPro?
     
  2. robbieduncan Moderator emeritus

    robbieduncan

    Joined:
    Jul 24, 2002
    Location:
    London
    #2
    a 2Ghz Core2Duo CPU is much faster than a 2Ghz G5.
     
  3. DoFoT9 macrumors P6

    DoFoT9

    Joined:
    Jun 11, 2007
    Location:
    Singapore
    #3
    yea thts a kinda wierd comparison. id like to see the results tho if a test is done.
     
  4. mrfaize thread starter macrumors member

    Joined:
    Sep 16, 2007
    #4

    I couldn't find out if Universal Binary apps are more efficient on Intel than PPC or not. I can't afford a MacPro so was wondering if a G5 would be quicker than an Intel Mac (iMac / Mini). I won't go for a iMac as I have a screen which I want to continue to use occasionally on my existing G4 PowerMac (OS9 apps) .

    So that leaves a maxed Mini or a second user G5 which are about the same in money.
     
  5. Umbongo macrumors 601

    Umbongo

    Joined:
    Sep 14, 2006
    Location:
    England
    #5
    The Core 2 Processors are faster, not by astonishing amounts, but they are. With a mini of course you'd be hampering yourself with a slower hard drive. If you can get a good price on a G5 then they are still worth looking at.
     
  6. mrfaize thread starter macrumors member

    Joined:
    Sep 16, 2007
    #6
    So Universal Binary apps run faster on Intel then? Is there no benefit in the G5 architecture?

    A Mac Mini C2D 2.0ghz will be quicker running UnivBin applications than a G5 2.3Ghz ? If so it's a no-brainer but was just trying to be thorough with my research.

    Thanks
     
  7. robbieduncan Moderator emeritus

    robbieduncan

    Joined:
    Jul 24, 2002
    Location:
    London
    #7
    You can see some comparisons here. A CoreDuo smokes a G5 (almost twice as fast in their iMove HD test at the same clock speed). Core2Duos are a bit faster again.
     
  8. mrfaize thread starter macrumors member

    Joined:
    Sep 16, 2007
    #8
    Was just having a dig around on macspeedzone.com also. Swaying towards a Mini over a G5 and maybe upgrade the HDD to something quicker.

    Edit: A maxed Mini from Apple is around £600 (with Apple memory!!) and a second hand 2.0Ghz G5 at our favorite auction site is about £600 - £650.
     
  9. mrfaize thread starter macrumors member

    Joined:
    Sep 16, 2007
    #9
    Those stats sort of speak for themselves as the apps were UB so clearly similar spec'd machines are quicker in the Intel guise.

    Edit: Here is also compare a C2D 2.0Ghz Mini to a Dual 1.8Ghz G5 and the Mini is faster so I think a new Intel is the way to go. Let's see what Steve gives us in January if anything.
     
  10. MacsRgr8 macrumors 604

    MacsRgr8

    Joined:
    Sep 8, 2002
    Location:
    The Netherlands
    #10
    Hmm.. Core Duo vs Single G5...

    IIRC a Dual Core G5 2.0 GHz and an Intel Core Duo 2.0 GHz weren't all that different.
    But you can be assured that future UB apps will rely more on SSE2/3 than on AltiVec (Velocity Engine) for faster SIMD instructions which will "speed' up the Intel compared to the G5.

    EDIT:
    There a more things to take into account. Raw CPU speed is one thing, but so are grfx, FSB and HD speed. You can get a Radeon X800 into a PowerMac G5, but the Intel Mac mini has to rely on the awful Intel GMA solution... The FSB of a Power Mac G5 Dual 2.7 GHz = 1350 MHz.... per CPU....
     
  11. mrfaize thread starter macrumors member

    Joined:
    Sep 16, 2007
    #11

    Point taken. The disk and graphics in the mini are not ideal (I can hold my breath to see if :apple: will do anything about it - probably buy in the New Year).

    FSB: considering my old G4 1.25Ghz MDD is 167mhz I think anything will be a step up, and would run CS3 better than my G4.

    However second user G5 Dual 2.7 sells at around £1000 (1400Euro $2000USD) and if I was to spend this much I'd bite the bullet and go for an entry level MacPro.

    Thanks for your comments.
     
  12. leekohler macrumors G5

    leekohler

    Joined:
    Dec 22, 2004
    Location:
    Chicago, Illinois
    #12
    I have a dual core 2.0 G5, runs CS3 and it's awesome. Get one if you can find one cheap. I love mine. I've had no desire to go Intel at all yet.
     
  13. mrfaize thread starter macrumors member

    Joined:
    Sep 16, 2007
    #13

    Cheap is the problem. In the UK second user G5 2.0Ghz go for around £650 $1300 (current exchange rate) on ebay. That's about half a new MacPro and the same as a new iMac obviously with warranties.
     
  14. DoFoT9 macrumors P6

    DoFoT9

    Joined:
    Jun 11, 2007
    Location:
    Singapore
    #14
    hhmm kinda same thing here.

    on ebay a dual 2.0ghz will cost around $1000-$1500 AUD. the new MP costs $2999
     
  15. Umbongo macrumors 601

    Umbongo

    Joined:
    Sep 14, 2006
    Location:
    England
    #15
    It'll be intertesting to see what happens to the current mac pro used prices when new ones come, I'm sure if you are persistant (and with a bit of luck) you could pick up a 2.66GHz system for a good price.
     
  16. Jiddick ExRex macrumors 65816

    Jiddick ExRex

    Joined:
    May 14, 2006
    Location:
    Roskilde, DK
    #16
    Universal Binary simply means that the application has been compiled with two kinds of source code, one for PowerPC and one for x86 architecture and then the OS will choose the correct one.

    Correct me if I am wrong.
     
  17. mrfaize thread starter macrumors member

    Joined:
    Sep 16, 2007
    #17
    I think you're correct but my original question was if either is more efficient (faster) on comparable hardware.
     
  18. Jiddick ExRex macrumors 65816

    Jiddick ExRex

    Joined:
    May 14, 2006
    Location:
    Roskilde, DK
    #18
    This will just be difficult to do, seeing as we are having different kinds of hardware as well. Since the g5 is much older technology wise than a Core2Duo chip, the Core2Duo will most likely show better results, even though the PowerPC code runs faster on the g5...
    Heck, even Office through Rosetta runs faster on my MBP than on my iBook... ^_^


    Ignore me if this point has already been made.
     
  19. Anonymous Freak macrumors 601

    Anonymous Freak

    Joined:
    Dec 12, 2002
    Location:
    Cascadia
    #19
    Alright. Lets say you have a dual-core, single socket Power Mac G5, and a dual-core, single-socket Mac Pro. (Yes, they didn't sell that model, but it's the closest you can get.) Both with Quadro 4500 graphics, with identical hard drives, both with 2x1 GB RAM sticks. This is as close as you can get to identical systems. (Since the Quadro is the only video card that is the same between the Power Mac G5 and the Mac Pro.)

    From the benchmarks at spec.org, the dual-core Mac Pro will win over the G5 on integer by about 66%, and floating point performance by about 15%. (I compared a 2.5 GHz PowerPC 970MP to a 2.66 GHz Xeon 5150, then extrapolated for the actual speed difference.) The CPU2000 raw scores are about 1600int/2100fp for the 2.5 GHz G5, and 2800int/2500fp for the 2.66 GHz Xeon. Unfortunately, the newer CPU2006 scores don't have any PowerPC chips in the database, only the higher-end POWER chips.
     
  20. mrfaize thread starter macrumors member

    Joined:
    Sep 16, 2007
    #20
    Thanks - the answer I was looking for. Intel all the way! Now wait for MWSF 08!
     
  21. diamond.g macrumors 603

    diamond.g

    Joined:
    Mar 20, 2007
    Location:
    Virginia
    #21
    Question:
    I was doing some research into that. That seems excessive considering it can only talk to the memory at 266mhz. Do PowerPC cpu's follow the same multiplier concept as x86 cpu's? I am trying to understand why the PM didn't use higher bandwidth ram, I mean seriously the system had 20GB/s of bandwith and was fed with memory that topped out at 4.2GB/s. That seems silly to me.
     
  22. DoFoT9 macrumors P6

    DoFoT9

    Joined:
    Jun 11, 2007
    Location:
    Singapore
    #22
    the numbers show that the intel's are faster, but i wonder what the actual performance will be.... i remember that the ppc g5's were incredibly fast, but i suppose the intel's will be faster.
     
  23. Anonymous Freak macrumors 601

    Anonymous Freak

    Joined:
    Dec 12, 2002
    Location:
    Cascadia
    #23
    The first-generation 1.6 GHz G5 used PC2700 DDR SDRAM (166 MHz, double-pumped to an effective 333 MHz,) every other 2003 and 2004 G5 (all of the PCI and PCI-X based models,) used PC3200 DDR SDRAM (200 MHz, double-pumped to an effective 400 MHz;) the 2005 G5s (the PCI Express based models) used PC4200 DDR-2 SDRAM (266 MHz, double-pumped to an effective 533 MHz.) Every G5 used "dual-channel" RAM, which means that the total available memory bandwidth is double that of a single module.

    This means that the 1.6 GHz G5s have 5.33 GB/s of memory bandwidth, the other PCI/PCI-X models have 6.4 GB/s, and the PCI Express models have 8.53 GB/s of memory bandwidth.

    For comparison, the fastest front side bus model, the 2.7 GHz dual-socket, single-core 2004 PCI-X G5 used a 1350 MHz front side bus (all except the 2004 1.8 GHz single-CPU model used a front side bus that was 1/2 the main processor speed. The 2004 1.8 GHz single-CPU model, and the iMac G5s, used a front side bus that was 1/3 the main processor speed.) The 1350 MHz front side bus has 10.8 GB/s of available bandwidth, per processor.

    This means, yes, most of the Power Mac G5 models do not have sufficient memory bandwidth to saturate even a single processor, much less two. This has been true of almost all 'consumer' level, and even most 'workstation' level computers for the better part of a decade and a half. (Basically since CPU multipliers started being used in the mid-to-late '90s.) The only system that can saturate its CPU bus is the 2004 single-1.8, which uses a 600 MHz CPU bus, for 4.8 GB/s, compared to its memory bandwidth of 6.4 GB/s. If they had used 400 MHz (effective) DDR on the 1.6 GHz model, then its processor bandwidth and memory bandwidth would have been equal. Every other model is bandwidth starved.

    Correction: The last-generation 2.0 GHz single-socket, dual-core model had 8 GB/s of processor bandwidth (since both cores used one front side bus, unlike the earlier models,) yet 8.53 GB/s of memory bandwidth.

    Oh, and no Power Mac G5 ever used 266 MHz RAM. (Well, technically DDR-2 533 is really running at 266 MHz, just "double clocked" to an effective 533 MHz.)

    By comparison, the Mac Pro uses 333 MHz (double-pumped to an effective 667 MHz,) DDR-2-based FB-DIMMs, in two or four channels, for a maximum 24 GB/s.
     
  24. GreatDrok macrumors 6502a

    GreatDrok

    Joined:
    May 1, 2006
    Location:
    New Zealand
    #24
    Really? I have a 933Mhz iBook G4 and a CD 2Ghz MacBook Pro and the iBook runs Office X much faster than the MBP does. I tried Office 2004 on the iBook and dropped it because it was way too slow and it was very slow on the MBP too but X is reasonable on both but definitely faster on the G4. Performance on the MBP is probably more like what you would get on a 700Mhz G4 IMHO.
     
  25. DoFoT9 macrumors P6

    DoFoT9

    Joined:
    Jun 11, 2007
    Location:
    Singapore
    #25
    yes tahts really interesting, that would be because the MBP has to use rosetta to run office.

    i find that my dad's ibook 1.2ghz runs office much much quicker than my mbp, and probably is around even if not quicker than my 2.8ghz imac :eek:
     

Share This Page