Sign in with
Sign up | Sign in
Your question

What have we learned about safe temps/voltz for Sandy Bridge?

Last response: in Overclocking
Share
September 3, 2011 9:00:13 PM

Hi all,

It's been nearly 8 months since Sandy Bridge was released. Is there an agreement on what constitutes safe voltages and temps for 24/7 use (today)? I've been researching the question and it appears there is a lot of disagreement, plus, many of the forums I've come accross thus far seem a bit outdated.

I just built a system with the following components and am wondering if my Vcore is safe for 24/7 use:

Intel Core i7 2600K (4.9GHz 1.50Vcore)
Gigabyte Z68 XPUD4 Motherboard
Dual MSI GTX560 Ti Hawk in SLI
Corsair AX850 power supply
Corsair H100
Corsair 600T-White
Corsair 16GB DDR3 (4x4GB) 8-8-8-24 1.5V
2 Crucial M4 128GB SSD's
1 Terabyte Samsung F3 HDD
Samsung Blueray Optical Drive
Scythe Fan Controller
5 120mm Fans
1 200mm fan

Thanks!
a b à CPUs
a b K Overclocking
September 4, 2011 2:57:18 PM

I think your at the upper limit if not the limit to what that chip can handle safely 24/7 for extended periods of time. I don't think there is actual documentation stating what the max voltage for the 2600K is. Only info on what the max VID range is, and they are not the same.
m
0
l
a c 100 à CPUs
a c 235 K Overclocking
September 4, 2011 4:54:15 PM

Intels max tested voltage is 1.520v listed in section 7:4 table 7:1, as for safe 24/7 use, that totally depends on the cooling solution being used and the constant temperatures daily achieved.

There's really no arrived conscientious solid data, simply because there are various ways to overclock the SB K series CPUs, my personal recommendations would be to stay below 1.40v for longevity, and spike it up for bragging rights, but the cooler you keep it the higher you can go with a 24/7 clock, so it still varies between users.

1.50v seems on the high side for 4.9Ghz, what are your temps running, I don't see any temperature references in your post?

I run my 2500K 24/7 at 4.5ghz @1.325v, simply because it can do everything I need it to at that level, I have the cooling capability to run it 5.0ghz 24/7, but what for?, it's totally not necessary for any applications I'm presently running.

So why stress and shorten the life of the CPU for no consequential gain needs, it's kinda like running a race car aggressively on a track constantly, but you're the only one racing, what's the point.

Bottomline; the cooler you keep it, and the less voltage you use to find a 24/7 level you're happy with, the longer your CPU will last, because once you cross the line and damage your CPU, it will never be the same again.

Once they're damaged from the voltage and heat, performance wise they're done.
m
0
l
Related resources
a b à CPUs
September 4, 2011 5:49:39 PM

4Ryan6 said:
Bottomline; the cooler you keep it, and the less voltage you use to find a 24/7 level you're happy with, the longer your CPU will last, because once you cross the line and damage your CPU, it will never be the same again.

Once they're damaged from the voltage and heat, performance wise they're done.


Does the CPU still work after that and if yes, how would you notice that they're "done"? How big is the performance drop?
m
0
l
September 4, 2011 6:10:03 PM


Thanks for your feedback! My temps are included in the following IMG link: " alt="" class="imgLz frmImg " />

Plus, I'm utilizing Intel's power saving features when the CPU is in idle mode; I'm not sure how much of an improvement that is, because the lowest the VCore drops is to 1.46V when in idle, at least that's what EasyTune6 is reporting.

I think I got a CPU from a bad batch, as one of the Sandy Bridge builds I put together in May 2011 only required 1.42V for 4.9GHz; whereas, this CPU needs 1.5V for 4.8GHz!

I see what you mean about running at full power all the time and agree that it's not neccessary.

Do you think my temps are decent enough to run this CPU at 4.8GHz 24/7 while also using Intel's power saving features when in idle mode?

Thanks!
m
0
l
a c 100 à CPUs
a c 235 K Overclocking
September 4, 2011 9:34:08 PM

Those temps are kinda on the toasty side after 3 minutes of Prime95, it usually takes 20 minutes to get in the max heat range with small FFTs.

Try using Real Temp instead of Core Temp, Real temp works better with Intel.
m
0
l
a c 172 à CPUs
a c 197 K Overclocking
September 6, 2011 8:11:10 AM

4Ryan6 said:
Intels max tested voltage is 1.520v listed in section 7:4 table 7:1, as for safe 24/7 use, that totally depends on the cooling solution being used and the constant temperatures daily achieved.

I wondered about that the first time I saw the "1.52 volts" figure in the Intel docs.

65 mn CPU recommended limits: 1.50 volts. 45 nm limits: 1.45 volts. I would have expected the 32 nm limits to be around 1.40 volts.
m
0
l
a c 100 à CPUs
a c 235 K Overclocking
September 6, 2011 3:40:24 PM

^ Logically I totally agree!
Somehow, this SB line is capable of handling it?
m
0
l
September 13, 2011 1:02:52 AM

4Ryan6 said:
Those temps are kinda on the toasty side after 3 minutes of Prime95, it usually takes 20 minutes to get in the max heat range with small FFTs.

Try using Real Temp instead of Core Temp, Real temp works better with Intel.


I ran Prime95 for 6 hours; everything was fine; no crashes. However, I decreased the speed to 4.6GHz and 1.45Vcore. I think I just have a bad CPU. Don't get me wrong, it works perfectly, just wish the stock Vcore was lower. My max temp is 65c at 100% CPU utilization on RealTemp. Also, my temps were the same when comparing CoreTemp to RealTemp; there was probablly a 1 degree difference.
m
0
l
!