Bare Feats Speed Tests: G4 not so bad after all?
From here: <a href="http://www.barefeats.com/pentium4.html"; target="_blank">Bare Feats speed tests</a>
Summary: G4 didn't win, but neither did it get smoked in a variety of real-world tests concocted by author. His methodology, while not "scientific" or comprehensive, seems reasonable (others can probably judge it better) at least. Even when running non-Altivec code, the 1 GHz DP Mac was very close to the 2 GHz SP P4. I, for one, was impressed.
Given all the whining and bleating about how Apple is SO incredibly slow compared to Wintel hardware around here, I was surprised at the results. I think we should all congratulate Apple Engineering for extracting so much performance out of such dated hardware. They've obviously worked very hard to remove bottlenecks and balance their systems so that every subsystem runs at close to its peak efficiency. Current Macs may not be the fastest stuff on the planet, but they certainly appear competitive.
What's this got to do with Future Hardware? Well, as I said earlier, if Apple can coax this much performance out of 3-4 year old technology, it ought to be unbelievable what things will be like when they finally make the jump to DDR, HT, RIO, the "G5" (whatever it ends up being) and the rest of the new-tech buzzwords. Whether the new stuff comes out later this month or next spring, it's going to be worth the wait, IMHO.
Any insights/opinions about this guy's testing and its implications for Apple's future? Or am I letting myself get carried away here? <img src="confused.gif" border="0">
Summary: G4 didn't win, but neither did it get smoked in a variety of real-world tests concocted by author. His methodology, while not "scientific" or comprehensive, seems reasonable (others can probably judge it better) at least. Even when running non-Altivec code, the 1 GHz DP Mac was very close to the 2 GHz SP P4. I, for one, was impressed.
Given all the whining and bleating about how Apple is SO incredibly slow compared to Wintel hardware around here, I was surprised at the results. I think we should all congratulate Apple Engineering for extracting so much performance out of such dated hardware. They've obviously worked very hard to remove bottlenecks and balance their systems so that every subsystem runs at close to its peak efficiency. Current Macs may not be the fastest stuff on the planet, but they certainly appear competitive.
What's this got to do with Future Hardware? Well, as I said earlier, if Apple can coax this much performance out of 3-4 year old technology, it ought to be unbelievable what things will be like when they finally make the jump to DDR, HT, RIO, the "G5" (whatever it ends up being) and the rest of the new-tech buzzwords. Whether the new stuff comes out later this month or next spring, it's going to be worth the wait, IMHO.
Any insights/opinions about this guy's testing and its implications for Apple's future? Or am I letting myself get carried away here? <img src="confused.gif" border="0">
Comments
I ADVISE ANYONE READING THIS NOT TO BUY A NEW POWERMAC untill those machines depicted in spy pics around the net are offered for sale. They are being produced as we speak, DON'T be the IDIOT that buys a slow silver now. Apple deserves to lose at least a few million in landfilled Powermacs for the way they've treated their customers. Fvck'em!
The P4 is only running at 2 Ghz.. The Amd chips are overclocked from 1600+ to 1900+ and 2000+.
<strong>Testing the fastest G4 box against low-end X86..
The P4 is only running at 2 Ghz.. The Amd chips are overclocked from 1600+ to 1900+ and 2000+.</strong><hr></blockquote>
No...the fastest G4 box would be the Dual Proc Xserve.
[ 08-02-2002: Message edited by: hmurchison ]</p>
<strong>Barefeats...
Eugene, I'm sure I'm not the only one wanting to know the context of your comment. Is Barefeats fabricating these results in order to make Apple hardware look better? In any case, I have to agree that anyone who buys QSs at this time is making a terrible investment. Unless, of course, Apple can't offer much more than 1.2GHz at the high end in this next revision. In other words, if the next machines fall far short of expectations, then a heavily discounted QS sounds like a good deal. Apple better deliver soon, I believe that's the general consensus.