Interesting if true, cruel if not!
For what it's worth, it's not whether these numbers sound credible, more whether they sound incredible.
I remember the shifts from 8-bit to 16-bit and 16-bit to 32-bit, and the step changes in performance that those generational leaps delivered. On that basis, the Mac-doobe-doobe-do numbers seem relatively credible.
But what we need to remember is that of these are benchmarks based on first-generation hardware (nothing faster than a 2.0GHz), developer release operating systems, and non-optimised applications.
What would the numbers really be on 2.8GHz, 0.09 micron fabrication, and an optimised OS and applications?
Even more interestingly, what happens if you line up a quad-9xx running at 2.8GHz?
If these numbers are true, I would be surprised if it takes longer than three years to convert 90-95% of the entire existing professional Mac community to 9xx, and more than five years to convert 75% of the domestic community.
Combined with iTunes Music Store and who knows what else, I'm only sorry I don't have a shed load of loose cash to "invest" in AAPL (no I really mean invest, not give) - I wonder whether the Iraqis hid any cash in the south-east of England like the $600M that was found a couple of weeks ago.
I remember the shifts from 8-bit to 16-bit and 16-bit to 32-bit, and the step changes in performance that those generational leaps delivered. On that basis, the Mac-doobe-doobe-do numbers seem relatively credible.
But what we need to remember is that of these are benchmarks based on first-generation hardware (nothing faster than a 2.0GHz), developer release operating systems, and non-optimised applications.
What would the numbers really be on 2.8GHz, 0.09 micron fabrication, and an optimised OS and applications?
Even more interestingly, what happens if you line up a quad-9xx running at 2.8GHz?
If these numbers are true, I would be surprised if it takes longer than three years to convert 90-95% of the entire existing professional Mac community to 9xx, and more than five years to convert 75% of the domestic community.
Combined with iTunes Music Store and who knows what else, I'm only sorry I don't have a shed load of loose cash to "invest" in AAPL (no I really mean invest, not give) - I wonder whether the Iraqis hid any cash in the south-east of England like the $600M that was found a couple of weeks ago.
Comments
The benchmarks are false. Plain and simple. Apple's slowest 970 is not going to be equal to a P4 3Ghz on it's best day.
Originally posted by hmurchison
Ohhhhhh sooo tempting.....I.....B......argh I can't do it!
The benchmarks are false. Plain and simple. Apple's slowest 970 is not going to be equal to a P4 3Ghz on it's best day.
Maybe, maybe not!
Is a 1.0GHz Itanium II faster than a P4 at 2.4GHz? If not, why bother buying them? Extending that logic, why should'nt a 1.4GHz 9xx not beat the pants of a P4 running at twice the clock speed? After all, it's not the length of the clock cycle that matters, but what the entire system manages to accomplish in that clock cycle that's important.
Also, let's remember that the tasks benchmarked could be logically expected to make use of Altivec, that Altivec is a well respected SIMD unit, and that Altivec on a 9xx maybe a quantum leap away from Altivec on a G4.
It's not that I believe or otherwise, it's just it doesn't seem so incredible that I would dismiss it out of hand. I know that as a community we are used to disappointment, but I can't help but think that this might be real. After all, this could be a case of "crying wolf" but in reverse.
PS: Mods, I apologise whole heartedly for starting a new thread, I meant to put this in the relevant thread, but obviously hit a wrong button somewhere. My fault for being an insomniac.
I just "reeeeeally" don't want to get all excited and then be disappointed. Why hasn't Apple shut these benchmarks down?
Waiting for the fabled "Removed at the request of Apple Legal"
We really haven't seen what Altivec can do since it's been pretty bandwidth starved. The benchmarks are possible but where would they be coming from. Apple's been pretty good with keeping stuff locked down but MacBoudille has been making it seem like they have Steve on their payroll.
Originally posted by Mark- Card Carrying FanaticRealist
For what it's worth, it's not whether these numbers sound credible, more whether they sound incredible.
[offtopic]
pssst...
I used to live in West Byfleet, Surrey for the better part of 2000-2001.
I miss England.
[/offtopic]
Weren't the SPEC benchmarks out for the 970? I believe the 1.5ish GHZ 970 is a shy lower than the P3 3GHZ...
Originally posted by chych
Weren't the SPEC benchmarks out for the 970? I believe the 1.5ish GHZ 970 is a shy lower than the P3 3GHZ...
Exactly. The 1.25GHz G4 spec scroes are about 3-4 times less than a 3.0GHz P4. And a G4 with altivec can somewhat be close in some tasks to a Pentium (we have seen the "bakeoffs"). So if the 970 and the Pentium 4 have similar spec scores now, why woudln't a 970 with altivec spank the shit out of a pentium 4?
Not to say the benchmarks are true or false (personally I won't believe it until Macworld posts something), but who knows. In another thread (on a different board) someone claimed that the benches are fake because the scores for the G4 and the P4 are form a barefeet test. However, since barefeets post how they do their test (or so I have heard) perhaps the 970 tester tested what he had, and used the barefeets results to compare the 970 to something. Does that make some sense?
I only hope Apple announces a 970 at WWDC because if they don't I'll have to listen to all the whiners again, but even if they announce one we will still get all the trolls complaining about the bogus benchmarks; being that these are totally false. It's a no win situation.