Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Performance Differences
- Thread starter Ryox
- Start date
- Sort by reaction score
Eidorian
macrumors Penryn
irishgrizzly
macrumors 65816
Hi Ryox, what kind of work are you doing? The real world performance difference may not be down to GHz speed, could be RAM or something else that will have the greatest influence.
Well, the difference between an octo-2.8 and a 3.2 will be enormous, and the difference between the 3.0 and the 3.2 will be huge.
And by enormous, I mean unless you're benchmarking and record the times taken by the 2.8 and the 3.2, you'll never notice the difference- to the point where someone might be able to convince you in a blind test that the 2.8 is faster with power of suggestion.
By huge I mean you'd have to peer pretty closely at the screen to see any difference in the numbers the two put up. I mean, come on. Do the math. 3.0 to 3.2? That's what? 6.6%?
Think about it this way... It's ten years ago, and you're buying between a 280, 300, and 320 MHz G3 (pretending they were running a bus speed of... I guess 80 MHz), and they all have the same cache size. Which is fastest? Not really a huge difference... particularly once you clocked your 280 to 340 MHz.
Fact is, you'll never notice the difference.. but if you can somehow figure out a way to overclock those systems, you'll be able to crank a 2.8 GHz Xeon a lot higher than 3.2 GHz, assuming you have a moderate amount of luck. Granted, dual quad-core Xeons aren't going to be the OCers that the dual-core Intel Core 2's are, but I'd bet they'll hit 3.6 GHz. You'd have to boot into Windows to do that, though, if it's doable. These Xeons are based on the (Penryn) Core 2 architecture- the only difference between them and Quad Cores (in addition to cache size), is, I think, the ability to have dual-CPUs (a tiny tweak).
And by enormous, I mean unless you're benchmarking and record the times taken by the 2.8 and the 3.2, you'll never notice the difference- to the point where someone might be able to convince you in a blind test that the 2.8 is faster with power of suggestion.
By huge I mean you'd have to peer pretty closely at the screen to see any difference in the numbers the two put up. I mean, come on. Do the math. 3.0 to 3.2? That's what? 6.6%?
Think about it this way... It's ten years ago, and you're buying between a 280, 300, and 320 MHz G3 (pretending they were running a bus speed of... I guess 80 MHz), and they all have the same cache size. Which is fastest? Not really a huge difference... particularly once you clocked your 280 to 340 MHz.
Fact is, you'll never notice the difference.. but if you can somehow figure out a way to overclock those systems, you'll be able to crank a 2.8 GHz Xeon a lot higher than 3.2 GHz, assuming you have a moderate amount of luck. Granted, dual quad-core Xeons aren't going to be the OCers that the dual-core Intel Core 2's are, but I'd bet they'll hit 3.6 GHz. You'd have to boot into Windows to do that, though, if it's doable. These Xeons are based on the (Penryn) Core 2 architecture- the only difference between them and Quad Cores (in addition to cache size), is, I think, the ability to have dual-CPUs (a tiny tweak).
The 5400 series penryn Xeons are clock locked ( no OC for you ) and the high end 3.2 runs extremely hot to the point it requires uprated cooling components vs the 2.8/3.0.
For the $800 increase from 2.8 to 3.0 you could buy 16GB of ram. that would make a much bigger difference.
http://eshop.macsales.com/shop/memory/Mac-Pro-Memory
http://eshop.macsales.com/shop/memory/Mac-Pro-Memory
with a 1600MHz front side bus...
wouldnt it be better to get the 3.2ghz because it is a whole multiplier with 1600?
wouldnt it be better to get the 3.2ghz because it is a whole multiplier with 1600?
Umbongo
macrumors 601
While I will agree that for most people they will never notice the difference, if you are doing hours of processing a week, every week, for a few years 3.2GHz over 2.8GHz can definatly save alot of time. I mean if you are waiting on processing 10 hours a week for 3 years the speed difference could save you almost 200 hours. It really depends on the type of work you do though.
newtech said:
There was some recent information on overclocking mac pros, it was based on how high they could push the memory, think they overclocked 3GHz duals to 3.4GHz, do maybe with the very fast FB-DIMMS they have for skulltrail maybe someone will overlock the new MP by quite a bit. Saying that though, Intel expect to have faster Xeons anyway, and the current ones have been overclocked to 5GHz with phase cooling (4.4GHz with water) already I think.
Overclocked Core 2 Duos/Quads ( 800 FSB ) are a far cry from Xeons ( 1333/1600 FSB ). 5400 series, aside from being clock locked, already are at the max FSB of 1600 that the 5400X chipset supports. If 5400's could clock higher intel would be only too glad to overcharge you even more.
Eidorian
macrumors Penryn
If you're going to need such a ridiculous amount of computational power where it'll take a workstation that long, I hope you have a grid or high performance computing network to work with instead.
Overclocking a cheap processor is fun but that's for a hobby enthusiast. We're talking about a workstation here.
Overclocking a cheap processor is fun but that's for a hobby enthusiast. We're talking about a workstation here.
ReanimationLP
macrumors 68030
The actual number the multiplier is timed by is 400 MHz.
So, 7x400 = 2.8 GHz, 7.5x400 = 3.0 GHz, 8x400 = 3.2 GHz
Its a quad-pumped bus. Like the RAM is 800 MHz, Dual Channel. 800 MHz RAM is actually 400 MHz too, since its DDR. 400 x 2 = 800 x 2 = 1600MHz.
For instance, I have a Pentium 4 with an 800 MHz bus.
My RAM is 400, in a Dual Channel configuration. So, RAM 200 x 2 = 400 x 2 = 800.
The P4 is a quad pumped bus of 200 MHz, so 200 x 4 = 800.
hmmm quite a discussion going on here.
I need a good processor currently I seem to be maxing out the processors on my Dual G5 2.3Ghz with 8GB RAM.
Now I have some cash on hand to upgrade that system and I want the best I can get so will probably go for the 3.2. Although I wont necessarily be upgrading straight away I'll upgrade when I need to.
I was interested in the performance differences between the processors speeds for such a high price.
I'll be using Adobe CS3, Logic Studio and Final Cut Studio
I need a good processor currently I seem to be maxing out the processors on my Dual G5 2.3Ghz with 8GB RAM.
Now I have some cash on hand to upgrade that system and I want the best I can get so will probably go for the 3.2. Although I wont necessarily be upgrading straight away I'll upgrade when I need to.
I was interested in the performance differences between the processors speeds for such a high price.
Eidorian
macrumors Penryn
It's not worth the high price. The prices are so much higher since it's more difficult to produce a higher clocked processor at that bus speed.
Quad 2.8 GHz x 2 is the best value.
There's a few discussions kicking around this forum about the new Mac Pro and Logic 8.
My call was to go for the 2.8, lots of RAM and good HDDs. What I'm reading online is confirming what I'm hearing in discussions and phone calls with folks working in Audio - the 2.8 is the sweetspot.
In the "old days", purchasing a processor with whole number multiplier was better in terms of performance. Should I just throw this old adage out?
Mr.Noisy
macrumors 65816
Performance Difference ?? What are you using the Mac For? indeed have you purchased one? ive a 2.66 quad with 8gb RAM (soon to be 10gb of RAM) for Photoshop it's the Bollox, so for performance increases % depends on the usage of your Mac, i will be buying one of the new generation 3.2ghz's as an upgrade in a year or so, but unless your video editing and so on, do you really need the power.
That adage never held water. Fractional and whole number multipliers are equally effective neither being better by any virtue other than absolute magnatude.
GregPacker
macrumors member