G5 x P4 x AMD 64 x Opteron
Bad news folks,
Take a look at this chart:
http://www.pcworld.com/news/article/...49,pg,8,00.asp
What do you think ?
n3o
Take a look at this chart:
http://www.pcworld.com/news/article/...49,pg,8,00.asp
What do you think ?
n3o
Comments
Originally posted by n3o
Bad news folks,
Take a look at this chart:
http://www.pcworld.com/news/article/...49,pg,8,00.asp
What do you think ?
n3o
I got scoffed at in another post by saying that Intel better get its butt in gear. That Athlon 64 is one kick-as* cpu. I've seen a number of publications where it beats the fastest Intel offering by more than 10%. In this industry, thats huge!!! Add in 64 bit capabilities and a beta version of Windows XP 64, and we're talking AMD PC revolution.
G5 holds its own against it. But IBM can't sit still (and won't). Converts will not come because of speed. They will come because Mac offers a superior overall experience.
Edit: I just went looking for it on their site but don't see it. I think I saw it in this months Laptop magazine. It must be coming soon.
But that's just me...
Originally posted by Jukebox Hero
G5 holds its own against it. But IBM can't sit still (and won't). Converts will not come because of speed. They will come because Mac offers a superior overall experience.
I agree 100%.
Well it goes without saying that all tests are biased, and I certainly wouldn't conclude that one chip is better than the other from this article (if you did, then you're missing out IMO). This article is just as PC biased as Apple benches are Apple biased. A few things to note:
Forget about architecture for a moment, and try to compare the two most comparable systems, that leaves you with the 1.8GHz G5 and the 2.2GHz A64 FX-51. Again, ignoring the architecture, just looking at clock speed the A64 should be about 25% faster. If you calculate it, for most areas where the A64 "beats" the G5, it does so by about 20-30%. That means it didn't beat the G5 at all. That leads me to believe that if either an A64 running at 1.8GHz or a G5 running at 2.2GHz were available (to at least even up the clock speed difference) then their would be little to no difference between the two chips. I almost missed the 2GHz A64 3200+. That one should be about 15% faster, and again, if you look at most scores (with the exceptions of the 2 below), it doesn't "outperform" the G5, it simply does about 15%, which it should, given that it's clock speed is about 15% faster.
The only areas where the A64 seems to really "beat" the G5 is the Quicktime under Premiere and the Quake. I can tell you PCs always do better on games, I won't argue about that. Why? I don't really know enough to say for sure, but they do have some advantages: They get better coding (NOTE: this isn't to say PC programmers are better, rather that there is a difference in creating something from scratch and porting something that was optimized for a specific OS), better hardware variety, better everything. Macs get a more limited selection of video cards and sound cards and the majority of games are ported from PC code. If the games were originally written for Macs so they took advantage of Mac's abilities from the getgo, then I'm sure the results would be different. In the end though, I can live with 141 frames per second in Quake
As for Word, well Word on OS X is a dog, I can say that from experience. So this "test" really doesn't prove anything.
I also liked how the article stated the Macs "aren't great values"; and their backup for that argument is that the dual-2GHz G5 is "about $200" more than a comparably equiped Alienware Aurora. When you're talking about buying a $3300-$3500 computer, $200+or- is not a big consideration. When you also consider the higher quality free software you get with the Mac (iPhoto, iMovie, & iDVD) compared to their free Windows counterparts, the Mac wins hands down.
In the end, this is fluff for the PC industry just as Apple's recent benchmarks are fluff for the Mac crowd.
Oh yeah, they "hand timed" these??? Not very accurate. I'd say these guys take after the MacAddict guys in their approach to "unbiased opinions" (if you have never heard of MacAddict, they are "pro-Mac" to the extreme) LOL!
In the end, the only useful info I come away with from this article is whether or not the architectural differences between Athlon and IBM (that is where the G5 comes from) really yield any advantage over one another. Obviously too many things come into play to really say whether the MB, RAM, what ever, contributed to the speed/slowness of a system, so it's tough to say. However when elminiating the difference for clock speed, they don't seem that different performance-wise...
So there's my uninformed opinion. Feel free to educate me, just be nice about it
Poor souls that are basing their opinion on these benchmarks, all I have to add.
Originally posted by fuzz_ball
Forget about architecture for a moment, and try to compare the two most comparable systems, that leaves you with the 1.8GHz G5 and the 2.2GHz A64 FX-51. Again, ignoring the architecture, just looking at clock speed the A64 should be about 25% faster. If you calculate it, for most areas where the A64 "beats" the G5, it does so by about 20-30%. That means it didn't beat the G5 at all. That leads me to believe that if either an A64 running at 1.8GHz or a G5 running at 2.2GHz were available (to at least even up the clock speed difference) then their would be little to no difference between the two chips. I almost missed the 2GHz A64 3200+. That one should be about 15% faster, and again, if you look at most scores (with the exceptions of the 2 below), it doesn't "outperform" the G5, it simply does about 15%, which it should, given that it's clock speed is about 15% faster.
So there's my uninformed opinion. Feel free to educate me, just be nice about it
Okay, allow me to educate if you will.
You are correct on the overall premise, these benchmarks don't mean a whole lot, but the paragraph where you talk about clock speeds is ill-informed.
For a moment, forget about clock speed. When comparing different families of chips it means nothing, absolutely nothing. So this talk of "should be about 15% faster" is wrong. Clock speed is only useful when comparing chips from the same exact family, P4s or Athlons or G5s are all different families, and their clocks speeds are not comparable.
Basically the point that needs to be made is that all of these applications, like Office Windows or Office Mac, share very little code, they share a name. In order to have a meaningful benchmark, you have to compile the same code on 2 different architectures without favoring either in optimization. Something like the SPEC benchmarks is intended to do just that. Comparing different applications (Office Windows versus Office Mac) tells very little about the speed of the processor. It does tell you that Office tasks are slower on the mac, which may or may not be important to you, but don't mistake it for indicating the speed of the processor
-Spyky
Originally posted by Spyky
Clock speed is only useful when comparing chips from the same exact family, P4s or Athlons or G5s are all different families, and their clocks speeds are not comparable.
-Spyky
That I actually do know, but thanks for driving home the point.
My original e-mail to my PC friend was a futile attempt to try and explain that there is more than meets the eye to this latest "chip shootout". Unfortunately, the fact remains, as much as we know that traditional benches that attempt to compare various architectures just don't work; people (publications/chip producers) will still do them. So as long as the public craves these (and they must, otherwise publications wouldn't continually churn them out), then it seems we must live with them for better or worse
<okay, here's my big rationalization> If the people cranking out the supposed "comparisons" are going to bend the rules in putting forth their info, then it is difficult to explain to someone what they're really seeing isn't true, unless I bend the rules myself in a lame attempt to explain what they think of as a black&white scientific conclusion, which it is not. Yeah, my explanation is about as solid as their comparisons, but if you're fighting fluff that is taken as fact, sometimes it's tough to get the point through
At least they didn't pull out the "scroll a Word document" benchmark.
Originally posted by fuzz_ball
That I actually do know, but thanks for driving home the point.
Okay, I'm glad you know that.
What is sort of humorous about the whole thing is that manufacturer's love to tout their MHz or MFLOPS or whichever irrelevant score if their processor happens to be on top. Then they have to backpedal and say things like "oh but MHz doesn't matter" when their clock speeds fall behind their competitor. It would be nice if they would just tell the real truth and stop making such a big deal about MHz. I think AMD did the right thing with going to their "performance numbers" and hiding the real MHz. But I think it is a safe bet that they will drop the current scheme and advertise real MHz if they are higher than Intel's offerings.
Intel has sort of painted it self in a corner at the moment, because they've been hailing higher MHz for years now, and the P4 is the ultimate result of that philosophy. It has a ridiculously large pipeline that allows it to achieve really high clock rates. However, their Itanium line will be unable to even come close to the same clock rates. Convincing people to ignore the P4 clock rate and buy Itaniums instead is going to be a tough sell.
But what do mac users care? We already have all our 64-bit goodness :-)
-Spyky
That will be more interesting than showing that Apple QT and Microsoft Word run best on their naitive platforms...