World's first overclocked Intel Dual Xeon Westmere benchmark

[Vinvin]

http://www.theinquirer.net/inquirer/rev ... enchmarked

With 2*6 cores (3.33 Ghz), it beats the 4* 6 cores AMD (2.8 Ghz) Istanbul at Whetstone and Drhystone !!
Incredible !
-> http://www.theinquirer.net/IMG/092/9409 ... 1268756764

[M ANSARI]

I think there is a new motherboard from EVGA that can push those chips quite a bit. I am actually thinking of getting such a setup, it should be fun to play with.

[ernest]

It says dual-core-xeon...

[LiquidNitrogenOverclocker]

It still won't beat my 4.5 GHz Intel Core i7-920 water-cooled radiator system:



and my 4.5 GHz Peltier-cooled system is 99.998% as fast:



Which you can read more about here:

http://www.liquidnitrogenoverclocking.c ... shop.shtml

[M ANSARI]

I am wondering, how does the Peltier compare to vapor? I know it is a lot less headache to setup, but what speed improvement are you getting over water?

[Vinvin]

Where do you see "dual-core" ?

It says dual-core-xeon...

[LiquidNitrogenOverclocker]

I am wondering, how does the Peltier compare to vapor? I know it is a lot less headache to setup, but what speed improvement are you getting over water?

The results of my "long" benchmark can be seen here:



which are on this page:

http://www.liquidnitrogenoverclocking.c ... l#page_top

The water-cooled unit was 19 seconds faster over a course of more than 3 and a half hours, so, basically, they are the "same speed". #5 on the list was water-cooled, and #6 was Peltier-cooled.

Peltier cooling is thermoelectric with no moving parts. Basically, if you join two different metals together to form a tight junction, and when you pass a current through the junction, a temperature gradient occurs. I have a better explanation of it here:

http://www.liquidnitrogenoverclocking.c ... l#page_top

Here is a recent CPU-z for one of the Peltier overclocks I did:



As for vapor, you might want to read about the 2-stage cascade systems one of my boxbuilders has worked on:

http://www.liquidnitrogenoverclocking.c ... l#page_top

These systems are much more complex, with pressure and temperature gauges that must be watched closely. There is a chance they could explode if the pressure soars out of control if unattended, so you have to be very careful with these implementations!

[ernest]

http://www.theinquirer.net/inquirer/rev ... ]dual-core[/b]-xeon-5680-setup-benchmarked

[M ANSARI]

They meant dual socket and this was obviously an error. With regards to Peltier I am aware of how they work but I tried it many years back and the units then were not too convincing because the heat gradient has to properly remove the "heat" part away. Probably the newer versions have that fixed. With regards to Vapor, I think a cascade is a little too much of a headache and probably not necessary to get good clocks. On a Quadcore the advantage of Vapor over water cooling is about 600 to 700 Mhz. I was interested in how the newer Peltiers compare to that.

[LiquidNitrogenOverclocker]

On a Quadcore the advantage of Vapor over water cooling is about 600 to 700 Mhz. I was interested in how the newer Peltiers compare to that.

The Peltier implementation I am doing is putting the plates in series, so that when it is in contact with the last plate, the cold side is close to -100 C. The advantage is the reduction in moving parts (water pump) and no worrying about leaks/seals breaking and flooding your computer. Also, "life" tends to sprout up inside your radiator tubes if you don't mix in some toxic chemicals with your water, always a pleasant surprise!

And even Obama likes my radiator-cooled 4.5 GHz systems, as this recent White House photo "proves"