CPU Question - Opinions Requested

They're basically the same chip. Perhaps the Xeon is binned a tad higher, but it smells of the PIII Xeons if you ask me- normal-line chip with the "Xeon" name sold for a markup. I'd go for the E6600, unless there are good reports of the Xeon 3060s overclocking much better than the E6600 if you're into that.
 

turpit

Splendid
Feb 12, 2006
6,373
0
25,780
Well,

Ive had my 6600 since july of last year, and other than TAT, and a few programs designed specifically to load up the cores, I have yet to do anything that has been able to task the CPU. The things I have done have been amazing. It will rip a CD faster than I could transfer the mp3s from and external drive, it doesnt even notice Norton2005 full system virus scan (it has actually run in the backround during a game without me even noticing it) It flies through renders on 3DSmax and while I have yet to run any of my video editing software, I suspect it process exceptionally fast.

Other than the fact it says zeon, the chip specs out the same in cache, FSB, socket and clock speed. Looks like a higher binned 6600 as MUEngineer said. If you want to OC to the max, spend the extra 10, if not the 6600 will overclock to 3.0GHz on air just fine. I havent even thought about overclocking.


<Edit> If you dont want to overclock, and are thinking about building an entire system, you may want to consider the X26000 @ $229. Its equal to or 1-2% faster overall than the E6600, depending on which review/test/blog you read/beleive.
 

Jake_Barnes

Splendid
<Edit> If you dont want to overclock, and are thinking about building an entire system, you may want to consider the X26000 @ $229. Its equal to or 1-2% faster overall than the E6600, depending on which review/test/blog you read/beleive.
Be kind of hard now since the GB DS3 board is arriving today :?
 

YO_KID37

Distinguished
Jan 15, 2006
1,277
0
19,280
Being efficient by a few degrees and maybe overall better Binned Chip it's better to spring over 10$ for that, It's almost like paying for Thermal paste, But your getting the thermal efficiency built in.

Plus, Everybody has a Core 2 Duo, why not just pay 10$ for the Xeon Premium bin and Brand Name :D
 

gOJDO

Distinguished
Mar 16, 2006
2,309
1
19,780
You'll pay $10 only because the box of the CPU has the "Xeon" label. Not worth IMO, because the chips are identical. The Ninja's review is not a solid proof that the Xeon runs cooler. Different CPUs will dissipate different heat. It depends on luck, the evenness of the surface of the CPU, how the thermal grease is applied, etc.

@turpit
The X2 6000+ is not worth the money, compared to both the E6600 or the 3060. It costs and performs roughly same, but dissipates heat and consumes power twice.
 
Stock benchmarks are going to be the same as both have the same FSB, are used in the same motherboards, same amount of cache, same clock speed, and same microarchitecture. The thing that might be illuminating is if somebody got about ten of each kind of chip and used the same board and cooler to do overclocking tests on them. My guess is that it wouldn't be a significant difference either way, but I'm not ready to pony up five grand for the CPUs to test.
 

yakyb

Distinguished
Jun 14, 2006
531
0
18,980
dont entirely agree with that if that $10 meant you could get a better psu or something save your money but then again if your budget is kinda loose the higher binning is probably worth it
 

Whizzard9992

Distinguished
Jan 18, 2006
1,076
0
19,280
<Edit> If you dont want to overclock, and are thinking about building an entire system, you may want to consider the X26000 @ $229. Its equal to or 1-2% faster overall than the E6600, depending on which review/test/blog you read/beleive.
Be kind of hard now since the GB DS3 board is arriving today :?

The difference between the XEON and C2D isn't so much a matter of performance as it is platform. I doubt you're going to see a real difference between the two in terms of performance.

In short, the C2D lets you overclock, but the XEON lets you double-up the chips (good upgrade path), and gives you access to things like PCI-X and motherboards with integrated SCSI/SAS. PCI-X lets you buy RAID cards for less money because the PCIe cards generally cost two times the amount of their PCI-X counterparts.

I went C2D because I like to overclock. I'm regretting it now because I want to upgrade to 15k drives, and the PCI-e controller is going to run me > US$300 alone.
 

fred_likes_fish

Distinguished
Nov 12, 2006
135
0
18,680
Its still only $10. To use your example, I'm sure you could pull another 10 from somewhere to cover that extra bit for the psu as well. If you can't, and your budget really is THAT tight, then you probably shouldn't be considering buying a new computer.
 
Whizzard, he's talking about the Xeon 3060 "Conroe," not the Xeon DP 5100 "Woodcrest" series. The Xeon DP 5000s let you put two chips together; the Xeon 3000 series are similar to the regular Core 2 Duos and run in the same LGA775 motherboards.
 

djgandy

Distinguished
Jul 14, 2006
661
0
18,980
Xeon sounds cooler imo :D

For 10bucks i'd probably take the xeon, server components makes you feel like they'll be that bit more reliable and durable.

That said the quality of core2duo is so good that unless you are pushing the clocks to the max you really wont notice the difference.
 
Its basically the same when it comes to benches. A few points here and there. The place the Xeon really shines is that it can run at .40V to .70V lower at the same speeds as the Core 2.
 
0.40 to 0.60 volts is a lot. IIRC, the default Vcore for top speed on a Core 2 Duo is 1.325 or 1.353 volts. 0.40 volts under would put it at 0.925-0.953 volts, which is little over the 0.85-volt idling speed. This is reasonable as I currently have my Athlon 64 X2 running at 0.12 volts over the idle voltage for top speed. I know, it's a different chip, but it still makes sense.

0.70 volts under that would mean 0.625 and 0.653 volts, which I'd have to see to believe. I believe that the current CMOS transistors in the chips have a band gap of about half a volt (Jack, correct me if I am wrong) so 0.625-0.653 volts would be getting down into the amplification region of the transistor, not the saturated region that digital logic requires. I'd have to look at the technical papers for the Intel 65 nm process to confirm.