Sandy Bridge EP Xeon E5 Lineup Revealed!

[Arachnotronic]

http://www.tomshardware.com/news/sandy-bridge-e-xeon-cpu-servers,13308.html

Enjoy!

 

[grkM3]

Xeon E5-2687W816 3.1 GHz 20 MB DDR3-1600 150 Watt ....yikes I dont know why they just dont go straight to 22nm with these

 

[ThatsABigOne]

150 watts for 8 core and 20mb of cache... I do not think that is too bad. We got cooling that can take care of 200 watts, even 220 watts.

 

[grkM3]

150 watts for 8 core and 20mb of cache... I do not think that is too bad. We got cooling that can take care of 200 watts, even 220 watts.

I dont know what cooling you have but my 2500k with 2 radiators with 4 fans hits 85c at 140 watts and 5 ghz

you wont get 4ghz on air cooling with these chips,they are 150 watt at 3.1 ghz.

Intel said there 22nm will cut power by 50%,thats when ill go to socket 2011 when they have ivys on it with lower tdps

 

[frostedflakes]

I dont know what cooling you have but my 2500k with 2 radiators with 4 fans hits 85c at 140 watts and 5 ghz

you wont get 4ghz on air cooling with these chips,they are 150 watt at 3.1 ghz.

Intel said there 22nm will cut power by 50%,thats when ill go to socket 2011 when they have ivys on it with lower tdps

SB-E is also a larger die. The same amount of power dissipation spread over a larger surface area -> lower temps. The power dissipated per unit area will be much lower than your 2500K @ 5GHz, you can't really compare the two based on power use alone. At least that's my understanding.

 

[grkM3]

SB-E is also a larger die. The same amount of power dissipation spread over a larger surface area -> lower temps. The power dissipated per unit area will be much lower than your 2500K @ 5GHz, you can't really compare the two based on power use alone. At least that's my understanding.

These will run HOT just like the older 1366 chips,the larger area has little to do with the tdp,if these 8 core pull 150 watts at 3.1 ghz it will blow way past 250 watts at 5ghz,good luck trying to cool it.

I guess you have never tried to over clock a high tdp chip,its not as easy to clock high.

I just ran my 2500k at 4.7ghz and it pulled 130 watts and hit 81c(intel burn test)with avx on with my water cooled setup.

these will hit 150 watts at 3.1ghz so even with my very same cooling I will hit higher than 81c testing the same way at 3.1ghz

now intel says on paper that there 22nm will cut power by 50% that is a HUGE win for us overclockers and will allow these to break 5ghz with 8+cores,heck maybe even 5ghz with decent air cooling.

150 tdp would be 75 watt with intels 22nm...Dont hold your breath hoping these 150 watt chips will clock high on air.

 

[frostedflakes]

Well area is important, like I said all else being equal, a 500mm^2 chip dissipating 150W should run cooler than a 250mm^2 chip dissipating the same amount of heat.

But they will be hot, power hungry chips no doubt. But it's not like anyone will be overclocking them anyway, they'll be used in climate controlled data centers where heat isn't an issue.

If you're saying that desktop SB-E won't be able to clock as high, though, you're probably right. Heat will definitely be an issue just like it was for Bloomfield.

 

[Arachnotronic]

These are Xeons based on the SNB uArch. I doubt they will be multi-unlocked so overclocking won't really be an option. These are server chips.

 

[grkM3]

I know these are server chips but they are socket 2011 so its basically a locked version of the up and coming sandy e chip,no wonder the rumors are that intel is not even going to include a heatsink with them.Most high end gamers are going to use extreme cooling for them.

I honeslty dont see anyone from a 1366 6 core system upgrading to socket 2011 until they put out the 22nm chips.

The more I think about it the 180 watt rumors might be true if they put out a high clocked 8 core chip and there was a 10 core es chip floating around so I wonder what that suckers tdp was.

The only reason they are going 32nm on these is becasue its there high end server market and they need a proven die shrink and cant risk going 22nm first without testing waters on the low end first.

I just was suprised intel is putting out a 150 tdp chip when they could of went 22nm and cut the power by almost half.

 

[notty22]

150 watts for 8 core and 20mb of cache... I do not think that is too bad. We got cooling that can take care of 200 watts, even 220 watts.

Agreed

 

[john3850]

Sandy Bridge EP Xeon E5 Lineup
A larger die has more contact area between the cooling block and chip which equal better cooling.
On my sb less then half wb is contact with the chip.

 

[imaheadcase]

Who cares how hot it runs, its the performance you want to know about.

 

[Nemesis 1]

I dont know what cooling you have but my 2500k with 2 radiators with 4 fans hits 85c at 140 watts and 5 ghz

you wont get 4ghz on air cooling with these chips,they are 150 watt at 3.1 ghz.

Intel said there 22nm will cut power by 50%,thats when ill go to socket 2011 when they have ivys on it with lower tdps

Just from the spec. you gave for your watercooling . What you have isn't much better than a good air setup . . If your cooling the gpu with 2 rads and 4 fans . It may not be as good as the best aircooling . I have a 2600k on water 5GHZ . its never come close to 85c. Thats with 2 rads 10 fans. Dead silent.

 

[Nemesis 1]

Who cares how hot it runs, its the performance you want to know about.

Ya intel will likely have to price them just under the godly AMD BD. Did I here someone say that SB ks were highend . Depends I guess on what you compare it to . You could call sb k highend if you owned an AMD cpu . Ya SB k would be high end , But if your talking with intels owners most would call SB high med 2600k Med 2500k/ 2600 Everthing below these would be high low end down to low lowend

 

[Vesku]

Has the die size been revealed yet? It will be interesting to compare clock/size/heat characteristics of these chips to the 2600K and BD (once it finally appears). Hope photonic research is progressing well for the fabrication companies, heh.

 

[grkM3]

Just from the spec. you gave for your watercooling . What you have isn't much better than a good air setup . . If your cooling the gpu with 2 rads and 4 fans . It may not be as good as the best aircooling . I have a 2600k on water 5GHZ . its never come close to 85c. Thats with 2 rads 10 fans. Dead silent.

You have a 2600k though,my 2600k did 5ghz at 1.395 volts and my 2500k takes 1.475 to hit the same clocks.

Just for fun can you run 5 passes of intel burn test 2.51 with 8 threads and avx on at 5 ghz.

Id like to see how good your cooling is.

Ill bet anything that you come real close to 85c if you load the cpu with max memory and 8threads with sp1 and avx enabled.

 

[dac7nco]

I honeslty dont see anyone from a 1366 6 core system upgrading to socket 2011 until they put out the 22nm chips.

A lot of Westmere users will be going S2011; [H] has a whole sub-forum for people who just can't get enough iron in their system. What you WON'T see, is micro-ATX boards for this thing... although it'd be hilarious to see someone try.

Daimon

 

[BlueBlazer]

A lot of Westmere users will be going S2011; [H] has a whole sub-forum for people who just can't get enough iron in their system. What you WON'T see, is micro-ATX boards for this thing... although it'd be hilarious to see someone try.

Humongous socket plus quad memory channels, it will be quite a (super technical) feat to try/achieve (squeezing all that into a micro ATX board). The motherboard manufacturers did it with X58 anyway, so they may try this monster as well.

 

[chiddy]

A lot of Westmere users will be going S2011; [H] has a whole sub-forum for people who just can't get enough iron in their system. What you WON'T see, is micro-ATX boards for this thing... although it'd be hilarious to see someone try.

Daimon

If i remember correctly, Clevo has already been seen flashing a S2011 board around not too long ago that's destined to find its way into an X7200 replacement.:biggrin: