a challenge to all pc users who prowl these boards.

Discussion in 'Macintosh Computers' started by Mord, May 24, 2005.

  1. macrumors G4

    Mord

    Joined:
    Aug 24, 2003
    Location:
    Old York
    #1
    allot of people have been saying the g5 get's owned by intel/AMD and every benchmark they have seen has proved it, please if this is your opinion can you post some links to some benchmarks.

    the rules:

    no benchmarks comparing different apps i dont care that mac the ripper is 20% slower on a g5 than whatever you use on your pc is.

    dont even try to use Gflops as a measure of performence.

    no overclocked benchmarks unless it's overclocked to the speed of a shipping cpu.

    when showing comparison calculate the cost per % of performance compared to the fastest rig, retail only no self builds.

    if in two weeks on tuesday there is not conclusive evidence that the G5 is slow i'm putting this thread in my sig and in every thread someone says "mac's are slow and they get owned by AMD/intel"

    keep it clean no flamming, that includes pointing out bad grammar/spelling.

    may the best cpu win.
     
  2. macrumors 68030

    cr2sh

    Joined:
    May 28, 2002
    Location:
    downtown
    #2
    Can you provide us with links to proper benchmarks? Name one or tell us where to find it... I'll run it on my Ahtlon 64 machine.... and my Dell Pentium M. :D
     
  3. macrumors 6502a

    feakbeak

    Joined:
    Oct 16, 2003
    Location:
    Michigan
    #3
    Why are you excluding self-builds? That's one of the advantages of x86 hardware - you can use the components you like and install Windows/Linux. Just because Apple keeps their system closed doesn't mean when you compare platforms you have to put Apple's contraints on the competition. If that's the case you have to put malware on the G5 before running your benchmarks. :D

    AMD/Intel systems can be self-built, that does not make them invalid for benchmarking purposes.
     
  4. thread starter macrumors G4

    Mord

    Joined:
    Aug 24, 2003
    Location:
    Old York
    #4
    i'm asking for the benchmarks, we have all seen barefeats and apple's benchmarks but allot of people quote mythical benchmarks where the g5 gets crushed.

    as for self builds, it's not fair for price comparison, i have a self build pc myself but the average joe dose not, for benchmark info self builds are fine just not for price comparison, and if they were they need licensed software.

    and cr2sh yours is a good rig to compare to a 1.8GHz G5, when someone here posts with a imac g5 or a 1.8GHz PM (dual or single you can disable a cpu with chud tools)

    and freakbeak seeing as there is no malware to speak of bar opener which just dose as it says, it opens, that would be a pointless exercise, i'm not denying that you can get a pc tower for less than a powermac or you cant get a 15" laptop with a faster cpu for less than a powerbook what i try to say is that for (exactly) what you get it's pretty good value.
     
  5. macrumors 6502a

    Blackheart

    Joined:
    Mar 13, 2004
    Location:
    Seattle
    #5
    One of the purposes of performing this benchmark is to figure the "price of performance" for the systems involved. Self-builds are not in the same category as Apple simply because Apple is a retail computer manufacturer. If comparisons are to be made to Apple hardware, with respect to price, then an adequate comparison would need to be drawn across the board. i.e. using retail manufacturers such as Dell and Alienware.

    Comparing a self-build here, would be much like building a car from scratch (using much of the same parts that a manufacturer would use) and calling it cheaper than retail because you were able to save yourself the money that the manufacturer would have stacked onto the original prices of the parts.

    Retail manufacturers offer a qualitative value to the products they sell. Apple, Dell, Alienware, etc. all offer warranty and service for their products. A self-build does not include this cost (don't think I'm going to sit here and argue the dollar value of this service for pages on end), so in order to make the comparison fair (and not nullify the value of the services provided by any manufacturer), every computer involved will have this service. Hence, no self-builds.
     
  6. thread starter macrumors G4

    Mord

    Joined:
    Aug 24, 2003
    Location:
    Old York
    #6
    note we are not bashing self builds, just saying it's unfair in a price comparison.
     
  7. macrumors 6502a

    feakbeak

    Joined:
    Oct 16, 2003
    Location:
    Michigan
    #7
    freakbeak? :confused:

    As for the malware I could whip up a little app for OS X that leaks memory for you! :D

    If you are doing price comparisons I can see why self-builds aren't fair. However, for a computer enthusiast that just wants the most powerful machine out there cost isn't going to be as big of an issue and being able to pick your components is a big plus - x86 wins hands down in that arena.

    I don't really care about benchmarks so much because there are far too many factors involved. It is way too easy to skew the numbers in favor of what you prefer. There are some lowel-level benchmarks that are straight-forward such as how long it takes to copy a large file or crunch through some floating point arithmetic. This may rule out factors such as the efficiency of the operating system libraries since these simple tasks can easily be coded in assembly. However, these types of benchmarks are not practical and do not really convey any useful information for daily use of your computer.

    Higher level benchmarking such as FPS measurements on games or rendering times in PS are more practical but the number of variables involved does not make it a fair comparison between hardware. Many applications are optimized for particular instructions sets (i.e. Altivec) and/or rely low-level libraries such as Quartz, CoreImage, DirectX, etc. Plus these apps also rely on the efficiency of the operating systems to manage resources, etc.

    When it comes down to it your experience and perception of performance hinges on many things other than raw computing power which is difficult to accurately measure in any meaningful manner. I tend to agree with Apple's mindset that is more about your "computing experience". There are things like I about Apple's hardware, OS and software and there are things I like about x86, Windows and Windows apps. Experiment with all the options and use what you like the most.

    Now, I am not saying benchmarks are completely useless. I think they do serve as a vague reference to how hardware performs, but you can't take it too seriously or think that a few points here or there makes one system better than another. If you use common sense and objectivity I think you can come to some reasonable conclusions. Apple is lagging behind in laptop performance these days. I know the G4's pack more punch per cycle but they are nearly 1 GHz behind and the FSB can only get up to 167 MHz - not a powerful combination. I don't think it's bad, but it's not comparable to what's going into high-end x86 laptops these days. The Power Mac G5 is doing well in my opinion. The only criticisms I have are the lack of PCIe and that for a huge case the expansion in terms of empty drive bays is rather pathetic. I think the new AMD Athlon 64 X2 and Intel Pentium D's will be giving the G5 a run for its money. Lets hope IBM can get dual-core G5's out within the next 6 months to keep pace.

    Hector, I understand you started this thread to try to quelch x86 fanboys who just come here to bash on PowerPC hardware. I am obviously not one of those people. However, I don't understand what any of these benchmark tests will prove. Numbers are numbers and can be manipulated towards one's desired hypothesis. My advice to computer users concerned with benchmarks is this: Try to obtain a good understanding of how hardware and software works together, use common sense and remain objective. Then, purchase the hardware/software that best fits your needs and that you enjoy using. Forget what other people say, it is irrelevant.
     
  8. macrumors member

    Joined:
    May 22, 2005
    Location:
    Newberg, OR
    #8
    I'd like to participate in this when I get my dualie G5, but unfourtunately I have a home-brewed system that isn't exactly the fastest out (it's an AthlonXP 3200+).

    HOWEVER, if I can get my paws on the Mac version of Photoshop CS, I'll pit it against the PC version of CS I have on this box for kicks, and run whatever other benchmark I can find that's the same on PC/Mac, then post my feelings and comparisons on the two different platforms.

    I really wish Bapco would make a Mac version of Sysmark, that'd be interesting.
     
  9. thread starter macrumors G4

    Mord

    Joined:
    Aug 24, 2003
    Location:
    Old York
    #9
    this thread is not out to prove the G5 is fast it's to prove that all these mythical benchmarks these people quote dont exist.

    as for self builds, if you license all the software they are not that cheap,

    just to state my opinion on x86-ppc comparisons.

    K8=G5=P4 of K8's PR rateing (approx)

    oh and no game benchmarks, they are heavily biased to the pc.
     
  10. macrumors 6502a

    csubear

    Joined:
    Aug 22, 2003
    #10
    http://www.gromacs.org/benchmarks/single.php
    Gromacs is a computational chemistry tool that uses Newtonian mechanics to simulate the movement of protiens.

    Its open source, and runs just about every where. It does not depend on a graphics card, just cpu and memory.

    The systems are not "up to date" but are very repersentive of what each arch can do. The only system on the line up that beats a dual 2.0 G5 is a quad 2.0 Opteron.
     
  11. macrumors 6502a

    csubear

    Joined:
    Aug 22, 2003
    #11
    Also notice that the G5 has the best pre-chip/pre-ghz ratios.(this is the column call rate) It only loses the a power3 and an alpha in this indicator.
     
  12. macrumors 68020

    Dreadnought

    Joined:
    Jul 22, 2002
    Location:
    Almere, The Netherlands
    #12
    Sounds a lot like Folding@Home, but you can't compair hte result of F@H of one platform with the other because of the optimized codes for windoos, unfortunately for the mac the code isn't completely optimized, but with every revision, they tweak it a bit.
     
  13. macrumors 6502a

    csubear

    Joined:
    Aug 22, 2003
    #13
    Not true. Its not quite folding at home, same concept, perhaps sharing some of the code, but F@H is solving the problem in a different way(distributed).

    I worked on a module for this project for a while and it is NOT optimized for windooos (click on the link and look at the OSs these machines are running, nobody in this field uses windows). It has a great deal of optimization BOTH SSE2 and Altivec.

    This app will show you the true horsepower a cpu has. This is something they run on supercomputers.

    If you are looking for numbers look at those on the gromacs site. It really tests processors, processors interconnects, and memory. There is very little disc IO.

    This app is a prime example of the reasons people are building super computers out of G5s. With a dual 2.0 g5 you can approach the performance of a 4 processor opertron box.
     
  14. thread starter macrumors G4

    Mord

    Joined:
    Aug 24, 2003
    Location:
    Old York
  15. macrumors 6502a

    Joined:
    Oct 7, 2003
    Location:
    London
    #15
    ok

    as a person who loves macs and g5's i would say apple's have their advantage's over pc's at some things but then so do pc's

    its about what programs you use and which os you as a person feel comfortable with not how fast the thing runs altough that is a factor which i would give to the pc many programs (gadget show etc) have proved it a mac is just not as fast but this is not the fault of the mac the software is to blame many devolpers lazy conversion of windows program for mac are not optimized for the mac like the pc but pure processing power goes to the pc of course when steve pops the 3ghz barrier then things will get cooking :)

    also i dont like apple's monopoly attitude at times
    with a pc a motherboard goes wrong you go anywhere pick up a replacement with a apple a logic board goes wrong kiss it goodbye or rip your arm off !!
     
  16. macrumors 68020

    G5Unit

    Joined:
    Apr 3, 2005
    Location:
    I'm calling the cops
    #16
    If you asked me, I would say that if a P4 ran on Tiger, it would be pretty darn fast.
     
  17. macrumors 68000

    Xtremehkr

    Joined:
    Jul 4, 2004
    #17
    From what I have seen, the PC Users just seem to hang out so that they can add negative rating to positive stories and positive comments to negative ones.

    There really aren't any trolls on the board.

    Though poor Lacero is constantly incurring the wrath of new people, and occasionally older ones too. Unintentionally though most of the time.
     
  18. macrumors 6502a

    Joined:
    Jul 20, 2004
    Location:
    California
    #18
    Frankly I'm pretty skeptical of most benchmarks as a general rule. Similarly configured systems to whatever I'm using at the time make claims of 40-50MB/s hard drive throughput. I run the same benchmark and see maybe 20-25MB/s, with some individual tests (like uncached writes) suffering at about 1MB/s. So except in identical systems with just one component switched out to make the comparison, I don't put much faith in benchmarks for Mac or for PC.
     
  19. macrumors 68030

    cr2sh

    Joined:
    May 28, 2002
    Location:
    downtown
    #19
    I disagree.

    However, I'm a mac user at heart.

    I wish there was a cross-platform standard for benchmarking.
    The one m,entioned above sounds nice... but I don't have a compiler and am not going to spend my night figuring out how to do it.
     
  20. macrumors 68000

    Xtremehkr

    Joined:
    Jul 4, 2004
    #20
    I was keeding. You have to wonder though, about who is behind the seemingly incoherent ratings at times. I know there are PC users here, I just don't understand why they are PC users. Same rule applies :D
     
  21. thread starter macrumors G4

    Mord

    Joined:
    Aug 24, 2003
    Location:
    Old York
    #21
    as i have said this head is not about debating which is better, it's just about proving the xeon/opteron are not way ahead of the G5 as most people say

    and i dont think pure processing power dose go to the pc if you go to barefeats when the benchmarks have been equally optimized the G5 dose well, just look at quake 3, it's so old it's had about every optimization in the book, it's the one game that can be used as a fair benchmark.

    what i want to see is if any of the pc users can pull these benchmarks because personally i think they are BS.
     
  22. macrumors newbie

    Joined:
    Mar 16, 2005
    #22
    Fairness is the main problem!

    I definitely believe that the G5 is truly the fastest desktop processor chip among all those AMDs and Intels out there, without a doubt.

    However, this takes into account that everything is fair. The problem is, very little is fair towards Macs. I give the following four reasons:

    1. It isn't fair that PCs can be self-built (saving much money), while Macs can't be self-built.

    2. It isn't fair that Windows has an over 90% user base, even though we have a better OS.

    3. It isn't fair that developers don't optimize their software on the Macs as much as they do on the PC.

    4. It isn't fair that games are considered a standard benchmark in most mainstream circles which, coincidentally, is the single weak point of the Mac.

    In short, Doom 3 is why many people think Macs are slow and overpriced. :p

    But seriously, the unfortunate fact is that all the unfairness means that benchmarks will usually go in favor of the PC. Unless you're Apple.
     
  23. thread starter macrumors G4

    Mord

    Joined:
    Aug 24, 2003
    Location:
    Old York
    #23
    or you do a fair benchmarks with real world application tests aka the stuff people acctually use there macs for.
     
  24. macrumors newbie

    Joined:
    May 4, 2005
    #24
    But one thing always baffles me about the economics of self-build ...

    Doesn't the build time have a monetary value as well???

    How much time / effort / money does it take just to locate all the relevant components, etc.


    I'm not a particularly high earner but I reckon my rates for building a computer would have to be somewhere between £15 - 20 per hour --- otherwise I might as well get on with my job & pay others to get on with theirs ...

    A couple of days work would soon wipe out any 'savings' ...
     
  25. thread starter macrumors G4

    Mord

    Joined:
    Aug 24, 2003
    Location:
    Old York
    #25
    it takes me about 15 mins to assemble, and then an hour to install stuff.

    the point is it's a minority that builds systems and it really dose not cost that much less when you buy all the software (which i and most people obviously dont).
     

Share This Page