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.
Xeon vs i7?
- Thread starter Macpropro80
- Start date
- Sort by reaction score
Possibly. Or, something is going on in that Dell result that we don't know about like too many background operations. The take home point is that there is virtually no performance difference between Nehalem i7's and Nehalem Xeons. How that silly belief got spread, I have no idea. I'm glad we now have the benchmarks to start putting that myth to bed. Both chips are based on the same design. The primary difference being the Xeon's ability to be used in multi-chip configurations.
Not necessarily.
Parts are refined as they continue to be manufactured, and as it's a new part, changes are more likely to occur. A change from a stepping of 4 to 5 is quite possible to make the parts more stable. It may also be a result of enabling the ECC function (again, stability).
Which BTW, some steppings have been noted in the past to be OC friendly for the otherwise same part.
I know the small slight revisions and errata fixes that stepping normally incurs but I remember specific mention that D0 was limited to just some silly labeling on the processor itself they wanted to correct.
I hadn't spent much time looking at the D0 parts. I already have a C0 part, and it's working fine. OC'd it to 3.93GHz, and it's stable. Temps hit 77C, so not too hot.
As for the Xeon 5500/3500 stepping differece between Core i7 parts, ECC support would by far make the most sense to me.
As for the Xeon 5500/3500 stepping differece between Core i7 parts, ECC support would by far make the most sense to me.
That and the QPI thing.
In reality, all Intel CPUs at the same level of technology, such as Penryn, Nehalem, come off the same production line. Bloomfield, Gainstown, and Beckton are pretty much the same. Intel intentionally disables some functions in low-end models. So it happened in the past that connecting two pins on the CPU actually made a UP CPU turn into a DP CPU.
We were referring to the stepping differences only.
Core i7 uses a stepping of 4, and the Xeon parts are stepped at 5.The E5520 (2.26GHz)l runs at 80W TDP, while the X5550 (2.66GHz) and X5570 (2.93GHz) run at 95W. The yet unseen W5580 (3.2GHz) runs at 130TDP. (The L55xx parts run at 60W TDP).
The W35xx parts all run at 130W TDP.
There's a convenient list here on Tom's.
From a Systems POV, it is.
Just scale & ECC functions primarily. You can add TDP here and there as well. 
Just for fun I thought it would be a good exercise to do a "real" comparison between a 2.66ghz Intel i7 rig and the Mac Pro 2.66ghz.
The reason being, it seems as though when quoting the cost of the Intel rig people intentionally quote what it would cost to build the machine with low end parts and no operating system.
The fact that the Mac has a top notch case, power supply, ECC RAM, etc, shouldn't be dismissed.
While this might not matter to the gamer that rebuilds their gaming rig every 12 months it is a big deal to someone who will invest in a high end machine and keep it running for 3-5 years.
So, without further ado, here's a more "realistic" comparison between the Mac Pro and a gaming i7 setup if the gaming setup used similar quality components, priced from Newegg;
Intel i7 2.66ghz Nehalem $288
OCZ Gold 3GB DDR3 RAM $120
Asus P6T Deluxe Mobo $300
1100 Watt 80+ Bronze PSU $200
Lian Li Aluminum Case $400
Slim Burner (combo drive) $50
Video Card $100
Aluminim Kb + mouse $50
Vista Premium 64 bit Ultim $180
640GB WD Black HD $80
So, if you build a PC that has similar quality components, etc, to the Mac Pro then you are looking at about ~$1800.
It then comes down to whether the ability to run OSX is worth the price difference.
And yes, you can build a far lower end PC by shaving corners on things like the case, PSU, and pirating the OS, but don't fool yourself that it's the same caliber box at that point.
The reason being, it seems as though when quoting the cost of the Intel rig people intentionally quote what it would cost to build the machine with low end parts and no operating system.
The fact that the Mac has a top notch case, power supply, ECC RAM, etc, shouldn't be dismissed.
While this might not matter to the gamer that rebuilds their gaming rig every 12 months it is a big deal to someone who will invest in a high end machine and keep it running for 3-5 years.
So, without further ado, here's a more "realistic" comparison between the Mac Pro and a gaming i7 setup if the gaming setup used similar quality components, priced from Newegg;
Intel i7 2.66ghz Nehalem $288
OCZ Gold 3GB DDR3 RAM $120
Asus P6T Deluxe Mobo $300
1100 Watt 80+ Bronze PSU $200
Lian Li Aluminum Case $400
Slim Burner (combo drive) $50
Video Card $100
Aluminim Kb + mouse $50
Vista Premium 64 bit Ultim $180
640GB WD Black HD $80
So, if you build a PC that has similar quality components, etc, to the Mac Pro then you are looking at about ~$1800.
It then comes down to whether the ability to run OSX is worth the price difference.
And yes, you can build a far lower end PC by shaving corners on things like the case, PSU, and pirating the OS, but don't fool yourself that it's the same caliber box at that point.
Remember the labor to assemble the box, and the skill required to do so. These are value-added functions that have a dollar value attached. And you'll never get the build quality that Apple's Mac Pros have.
Then run them side by side for 5 years and see which one still works. There is a reason why Apple leads the industry in customer satisfaction.
Plus firewire controller, wifi, bluetooth, and a sound card that supports TOSLINK input.
The Asus mobo should have onboard support for toslink optical and it has onboard sound.
Probably no wi-fi or bluetooth though, so you could add $100 for that.
Yes it should for the person that's trying to get identical or better PERFORMANCE! I'm seriously about to claw my eyes out that people cannot understand this. When a lot of people build hacks (myself included) they're not interested in using the same parts at all. What they're interested in is getting the same or better performance for less money. So, they buy cheaper CPUs, spend $100 on cooling, and OC the things. The fact that they aren't using a Xeon CPU or even a CPU that's natively clocked the same is completely irrelevant to their purposes. They simply want to encode as fast or faster for less money. Or, render a bunch of photoshop effects faster for less money. Or whatever.
When I built my quad core hack the quad core Mac Pro with less RAM, less HD space, weaker video card, less ports, and lower performance in every application cost at least twice as much. Did I cheat by OCing my system? Maybe. But guess what? I don't care! It's been running perfectly stable. And by the way, the cost of my system included the retail costs of two OSs; Vista Ultimate and OSX, and it was still less than half the cost of a lesser spec'd Mac Pro. I don't doubt that for some people the case, warranty, and avoidance of "hacking" make an authentic Mac Pro worth the extra cost. They're fine machines. But for other people the savings of building their own system is significant and the performance gained at less cost is even more significant.
Even granting the system you put together above, with very little OCing, that thing would blow the quad Mac Pro out of the water. So, for ~$1800 you're not just getting identical performance, you're getting much better performance.
I have been building and overclocking systems for 12 years. I'm very well aware of what it means to "chase performance" at all costs.
However, many of us eventually wind up in a place where this is not the #1 priority. We want good performance AND we want top notch fit and finish AND we want stability more than anything.
I've been at LAN parties watching peoples "rock solid" overclocked rigs BSOD/crash/lockup/etc.
I've also had to tear down, tweak and rebuild overclocked rigs because everything seemed to be stable until after a week or so when it was determined that system instability was causing problems like registry corruption.
The Mac Pro is no different than any other high end workstation. It is designed to operate quietly and reliably for 5 yrs. I've had lots of home build rigs and ALL of them have needed parts replaced in less time than that due to capacitors going out on motherboards, RAM going bad, etc.
This can happen on workstation class PCs but it's far less common. I have multiple workstation class computers at my wife's office and never have a lick of trouble out of any of them. Could I have built the same PC for hundreds less? Yes I could, but then I'd be dealing with all of the headaches.
Something that guys like yourself don't understand that makes some of us laugh is that some people just want a stable, reliable workstation that's built like a brick **** house and won't give them any trouble.
And they want to boot camp to play Crysis every once in a while.
I've been contemplating making the switch to a Mac for over a year now due to high dissatisfaction levels with Vista. Every time a PC fanboy gets wind of it they want to get into one of these debates, never understanding that I'm a working professional that simply wants a rock solid platform to work on that I can also use for occasional other uses such as running a Windows game.
I've built more than my share of hot-rodded rigs, overclocked to the max, and bragged about my benchmarks. In the real world, however, I actually need to do work. For hobbyists, hacking together a machine for bragging rights on speed versus cost is fun; and they get to pound their chests.
Apple doesn't give a crap about that market segment. It's so small that it makes no business sense.
Like you said, there is great value in having a stable, reliable, decent-performing rig. That's where the rubber meets the road for most of the computer-consuming public.
And something that guys like yourself don't understand is that some people can make those sorts of systems themselves. I've not had a single issue with any of my Hacks. Hell, I haven't even touched my wife's since I built it over a year ago. It just...works. I'm still using the A64 system I built at the end of 2003 and I haven't changed a single part inside the thing. I've never had to sacrifice stability for performance. Not once.
Look, there are perfectly good reasons to get a Mac. You even listed them. Not everyone wants to put in whatever time is necessary to build their own machine. To service it. Some people don't want the risk. Some people like a certain look; an aesthetic. All that costs money. But, at the end of the day, when we're each working on Photoshop projects or Final Cut renders, the fact remains that a person with a well built system is getting the job done just as fast or faster than the guy with the authentic system. And, he did it for less money.
Let me repeat. I'm not at all suggesting Macs are crappy. They're not. They're good looking and great performing machines. I love the OS. But for a capable subset of consumers, building can get them the same stability and better performance for hundreds and, in some cases, thousands less than the manufacturers.
I don't doubt this for a second. My point is simply that there are people that get stable, reliable, and decent-performing rigs for less than the manufacturers by doing it themselves. It's not for everyone. This isn't controversial.
Thousands is a serious stretch.
Also as someone has previously pointed out, they've yet to see the benchmarks from a real OS X application in which a Hack beat a Mac Pro.
If you want to build it cheaper, then do it, but you're wasting your time trying to convince everyone here what a genius you are. Many of us could build our own if we wanted to. Easily too.
You can also probably build a much higher performance car for less $$ than what a new one costs by getting all of the parts 2nd hand, from scrap yards, etc.
Doesn't mean I'm going to be convinced that its a superior ride.
What are you considering a "real" OSX application? Not that it matters. In systems with the same number of cores using the same chipsets, you'll get the same performance. I hope that's not surprising. Of course, the difference between the Hack and the Mac Pro, is that in the same scenarios, the Hack can easily be OC'd and thus, is able to perform better. Just have somebody with an i7 Hack run Cinebench. At the same clocks, their system will get the same performance a 2009 quad Mac gets. At higher clocks, it'll get better performance. No surprise. And unsurprisingly, this will be shown to hold just as strongly for octo-Hacks once we get DP motherboards.
You know what, you "win". We all bow down to you, as you are obviously far smarter than anyone stupid enough to spend money on a "real" Mac.
Wouldn't it be even better if everyone just ripped off OS X and ran it on their own hardware, putting Apple out of business?
Awesome man.
Wouldn't it be even better if everyone just ripped off OS X and ran it on their own hardware, putting Apple out of business?
Awesome man.