i5-2500k 5Ghz

bezant · Dec 17, 2011

Just bought an i5-2500K and after reading what others did to overclock, I've got mine on 5 Ghz easy and stable.

Max vcore is 1.424 and max temps are 86C. That's on air, and not very aggressive fans.

Here are the bios settings of my P8Z68-V, which are pretty much a copy from Sean Webster but with minor adjustments:

Ai Tweaker Ai Overclock Tuner: Manual
BLCK/PCIE Frequency: 100.0
Turbo Ratio: By All Cores
By All Cores: 50
Internal PLL Voltage: Enabled
Memory Frequency: 1333MHz
DRAM Timing Control: 9-9-9-24
EPU Power Saving Mode: Disabled

Ai Tweaker\ CPU Power Management >
CPU Ratio: Auto
Enhanced Intel SpeedStep Technology: Enabled
Turbo Mode: Enabled
Long Duration Power Limit: Auto
Long Duration Maintained: Auto
Short Duration Power Limit: Auto
Additional Turbo Voltage: Auto
Primary Plane Current Limit: Auto

Ai Tweaker (in the DIGI+ VRM section)
Load-Line Calibration: High
VRM Frequency: Manual
VRM Fixed Frequency Mode: 350
Phase Control: Extreme
Duty Control: Extreme
CPU Current Capability: 140%
CPU Voltage: Offset Mode
Offset Mode Sign: +
CPU Offset Voltage: 0.130
DRAM Voltage: 1.5
VCCSA Voltage: Auto
VCCIO Voltage: Auto
CPU PLL Voltage: 1.7
PCH Voltage: Auto
CPU Spread Spectrum: Enabled

Advanced\ CPU Configuration >
CPU Ratio: Auto
Intel Adaptive Thermal Monitor: Enabled
Active Processor Cores: All
Limit CPUID Maximum: Disabled
Execute Disable Bit: Enabled
Intel Virtualization Technology: Enabled
Enhanced Intel SpeedStep Technology: Enabled
Turbo Mode: Enabled
CPU C1E: Enabled
CPU C3 Report: Disabled
CPU C6 Report: Disabled

Hope that helps someone out there.

VirtualLarry · Dec 17, 2011

You may not want to keep your vcore at over 1.4v for long-term use. It is likely to degrade the CPU.

I know that a 45nm C2D E5200 degraded with over a year running at 1.425v.

dma0991 · Dec 17, 2011

bezant said:
Just bought an i5-2500K and after reading what others did to overclock, I've got mine on 5 Ghz easy and stable. Max vcore is 1.424 and max temps are 86C. That's on air, and not very aggressive fans.

You've got a recipe for an impending disaster. Do not assume that your CPU will be exactly identical to his and expect similar results without compromising something else.

Dadofamunky · Dec 17, 2011

I have somewhat more conservative settings for the voltage and mapped out the other settings the same way at 4.8 GHz. Running both LinX and Prime along with CoreTemp 1.0, I find that the system can quickly reach TJMax. However, once it does that, it starts downshifting the voltage and the clockspeed so that the chip never exceeds TJMax. Very interesting. No crashes at all, everything runs fine - except that at extended stress-test runs, the system throttles down the clock speed and voltage. I idle at 1.371V with these settings and sure enough, stress-testing yields exactly the same thing - an hour of Prime and I'm at 4.1 GHz, 1.3661V. Pretty high voltage for that minimal clock speed OC.

Stop all tests and it bounces back to 4.8 GHz, 1.3711V. I don't think I'll keep these settings but it's an interesting combination to try.

Since no one runs their systems on Prime95 or LinX to do anything useful, but I can't get the system to crash at these settings, my thought is that if you spend all your time running apps like I do (Photoshop, Illustrator, Framemaker, Madcap Flare, MS Office, various telnet and SSH sessions to routers and switches and such), this might be a potentially good setup combination for you to get your stuff running at 4.8 GHz or 5 GHz since none of that stuff stresses a modern computer out much.

But the others are right, keep the voltages down more. Not a good idea to go much beyond 1.36V with the SBs.

Dadofamunky · Dec 17, 2011

Meh. I stand corrected. I ran a couple of other diagnostics programs (CPU-Z, HWMonitor) and they both showed a shocking 1.5+ V. Completely unacceptable. Except for suicide benchies, I really don't think this configuration set is a good idea.

Arkaign · Dec 17, 2011

Going to ~4.7ghz with a bit lower volts/temps won't make jack squat difference in terms of the big performance picture, but will be a world better in terms of not cooking your $200 cpu with electromigration and/or your mobo with leaking caps in short order.

I realize 5ghz is a cool sounding stat for your system, but once you realize nobody much cares, dial it back. 22nm is imminent, and 5ghz *should* be a cakewalk safely with those CPUs.

BTRY B 529th FA BN · Dec 17, 2011

Screen shot?

Puppies04 · Dec 17, 2011

bezant said:
Just bought an i5-2500K and after reading what others did to overclock, I've got mine on 5 Ghz easy and stable. Max vcore is 1.424 and max temps are 86C. That's on air, and not very aggressive fans.

*Sits back and waits for the "My pc keeps bluescreening and it won't even stay stable at stock" post*

tweakboy · Dec 17, 2011

Those load temps of 86c is no good. The CPU is going to die in no time. But if you speed up the HSF fan then you will get a noisy OC.

I believe at 95c the computer will shut off itself. If your load temps are in 70;s c and 80's c then your CPU will have a short life span. Also try playing some games on that OC make shure its stable, but your cooling blows goat,, your brave... 86c yikes,,, My CPU shuts off at 85c ,,,,, Sandy is 95c I think,,,, thx gl

Puppies04 · Dec 17, 2011

As far as I know SB just clocks back at TJ max.

996GT2 · Dec 17, 2011

1.424V is fine if you have proper cooling, which you do not. 86C load temp is too high for 24/7 use. You want to keep the max load temp below the low 70s in stress testing programs if at all possible.

tweakboy · Dec 17, 2011

Yes, your Sandy will turn into Sad Bridge with those temps..... Have you tried any games with that OC ? See how it performs in real world and if it does not crash your comp. gl

fastamdman · Dec 18, 2011

Couple things. A, you don't want to have a high idle voltage. B, you don't want to have temps that high, usually max people like is 80c in LinX. C, you dont want to hit the thermal throttling point ever.

As far as running 1.4volts for 24/7 that is just fine but keep it at that point and under, nothing higher or you COULD experience degradation.

Most chips wont do 5ghz stable so if you are consider yourself lucky!

Schmide · Dec 18, 2011

Side note: 1 year 4 month lurk!

peonyu · Dec 18, 2011

996GT2 said:
1.424V is fine if you have proper cooling, which you do not. 86C load temp is too high for 24/7 use. You want to keep the max load temp below the low 70s in stress testing programs if at all possible.

Depends on what app takes it to 86c. If its just Linx I wouldnt worry about it, I hit 88c before I changed my cooling under that app on 1 core. But in any game or even video encoding my max temp was 77c. Im not saying high temps are nothing to worry about but benchmark apps are not very realistic [unless thats all you do]. My c2d system ran at a max temp of 78c for about 2 years and its still trucking away without a problem [it was a server setup]. Voltage is where it will kill your system the fastest. 1.4v is pushing it, 1.5 is just stupid.

Personally i would say 84c is the max on SB, these chips run hot and can take the heat from what Ive seen, older generation cpus could not handle anything above 70 and even older chips such as Athlon 64 bsod above 65c.

CaptainAx · Dec 19, 2011

bezant said:
Just bought an i5-2500K and after reading what others did to overclock, I've got mine on 5 Ghz easy and stable. Max vcore is 1.424 and max temps are 86C. That's on air, and not very aggressive fans.

tweakboy · Dec 19, 2011

If the temps are kept low the high voltage wont degrade the CPU.

For example Im close to 1.5v however my idle is 20's c and load 40's to 50's c ....

Intel probably recommends Sandy to be at 60's on full load. Anything higher and your degrading the processor.

If you have proper cooling you can up your voltage to whatever you want. However 78c or 86c is too much, lower your voltage and thus lower your OC.... gl

SeanJ76 · Apr 1, 2014

I've been running my 2500k@5.0ghz@1.460Vcore for over 3 years now, still going strong, highest gaming temps are 69-72C with a 95% load. Hyper 212 Plus on a MSI Z68A-GD-55 in a Antec 300 case(7 fans) :colbert:

http://www.3dmark.com/3dm/1947312 -Firestrike bench
http://www.3dmark.com/3dm11/6795820 -3Dmark11
Degradation does not occur from high voltages, degradation occurs from heat, you would have to be running above 90C for long periods of time to degrade the 2500k as the TJC max is 98C.(I've personally never seen mine go above 83C).

BonzaiDuck · Apr 1, 2014

SeanJ76 said:
I've been running my 2500k@5.0ghz@1.460Vcore for over 3 years now, still going strong, highest gaming temps are 69-72C with a 95% load. Hyper 212 Plus on a MSI Z68A-GD-55 in a Antec 300 case(7 fans)
http://www.3dmark.com/3dm/1947312 -Firestrike bench
http://www.3dmark.com/3dm11/6795820 -3Dmark11
Degradation does not occur from high voltages, degradation occurs from heat, you would have to be running above 90C for long periods of time to degrade the 2500k as the TJC max is 98C.(I've personally never seen mine go above 83C).

Geez. A Hyper 212+. I think I could find you another 5C drop in temperatures, and you could still keep the "green-bling-bling" effect.

There are some "seemingly authoritative" sources which suggest that electromigration begins to accelerate in a probability distribution above the lithography's so-called "safe upper-bound." But . . . it's "probability."

But . . . now . . . you got me all worked up. "I, Professor CHA-OS, will RULE the world when I buy my second 780 GTX card and some Green BitFenix Spectre-Pro 200mm LED fans! You will wither under the penetrating light of Professor Chaos' magic machine . . . !!. . . "

996GT2 · Apr 1, 2014

SeanJ76 said:
I've been running my 2500k@5.0ghz@1.460Vcore for over 3 years now, still going strong, highest gaming temps are 69-72C with a 95% load. Hyper 212 Plus on a MSI Z68A-GD-55 in a Antec 300 case(7 fans)
http://www.3dmark.com/3dm/1947312 -Firestrike bench
http://www.3dmark.com/3dm11/6795820 -3Dmark11
Degradation does not occur from high voltages, degradation occurs from heat, you would have to be running above 90C for long periods of time to degrade the 2500k as the TJC max is 98C.(I've personally never seen mine go above 83C).

Those are great temps from a Hyper 212+. My i7-2700K @ 5 GHz gets higher temps than that in LinX or Prime95, and I have a much bigger cooler. Granted, hyperthreading does increase temps, but I'm still surprised the temps are so low @ 5 GHz (and at your voltages) with a small heatsink like that.

I know you mentioned those temps are at 95% load; what are your full load temps in Prime95 or LinX?

sochee · Apr 3, 2014

Strong necro... only 3 years old..

videogames101 · Apr 3, 2014

yeah, anything below 1.45v is fine on SB imo, and 86C on stress testing? jesus if I could be so lucky with IVB

BonzaiDuck · Apr 3, 2014

996GT2 said:
Those are great temps from a Hyper 212+. My i7-2700K @ 5 GHz gets higher temps than that in LinX or Prime95, and I have a much bigger cooler. Granted, hyperthreading does increase temps, but I'm still surprised the temps are so low @ 5 GHz (and at your voltages) with a small heatsink like that.

I know you mentioned those temps are at 95% load; what are your full load temps in Prime95 or LinX?

Depends on how he measures his top load temperature. I'd say that 83C could be about right for processors in the SB-K category -- with that voltage and speed. He may be in the upper tail of the distribution, but those numbers seem correct.

The four core sensors, last I heard or read, have a tolerance or accuracy range of maybe +/- 6C. I believe that all the stories are bunk about "dark silicon" and how temperatures between cores under the same load vary. I think I can disprove it.

So you might average the data from four cores at a given point in time. You might average the maximum temperatures trapped by some software like CPUID HWMonitor -- but those would only be momentary peaks in a distribution. People who overclock may look at the highest temperature in the set and think that it is the only relevant temperature, when it's a likely and consistent error from a mean.

The paradox of this -- only speculating here -- the throttling temperature or Tj for a processor might be that for the high core if the sensors are used to trigger throttling. So once you knew which of your cores was the high one, you would temper your OC'ing efforts to reach a maximum of Tj - Ti if "i" is the hot core (so-called).

amenx · Apr 3, 2014

SeanJ76 said:
Degradation does not occur from high voltages, degradation occurs from heat...

Not necessarily. Electromigration can occur from voltages/current and degrade microelectronics and circuits.

Several years ago there was a wave of CPU degradations of the Wolfdale series (e8400/e8500), much of it reported and documented at xtreme systems. Many users did not have an issue with heat and kept their chips relatively cool, yet their CPUs degraded, including mine, despite it rarely going over 60c load temps.

BonzaiDuck · Apr 3, 2014

amenx said:
Not necessarily. Electromigration can occur from voltages/current and degrade microelectronics and circuits.

Several years ago there was a wave of CPU degradations of the Wolfdale series (e8400/e8500), much of it reported and documented at xtreme systems. Many users did not have an issue with heat and kept their chips relatively cool, yet their CPUs degraded, including mine, despite it rarely going over 60c load temps.

[I'm still wondering if what I noticed with my E8600 originated with the CPU or the motherboard. But after the first time I had to adjust the voltage, I stopped OC'ing it. And it's still ticking along at stock values in my bro's system.]

i5-2500k 5Ghz

Junior Member

No Lifer

Platinum Member

Platinum Member

Platinum Member

Lifer

Lifer

Diamond Member

Diamond Member

Diamond Member

Diamond Member

Diamond Member

Golden Member

Diamond Member

Platinum Member

Junior Member

Diamond Member

Member

Lifer

Diamond Member

Member

Diamond Member

Lifer

Diamond Member

Lifer