Yes! 5 GHz out of my 2500K!

[996GT2]

I've been running my 2500K @ 4.4 GHz, 1.292v for daily use, but today I went into BIOS and did a quick run to see how high it could go.

Needless to say I won't be running at this speed 24/7, but here's what came out of it:

 

[nyker96]

NICE ONE! what type of cooler?

 

[996GT2]

NICE ONE! what type of cooler?

Hyper 212+. Specs in sig.

Just shows that even with modest air cooling and a midrange mobo without any fancy "overclocking" specific features (like 12 phase power), you can still reach very high speeds with Sandy Bridge chips.

 

[OCGuy]

Wow monster chip!

 

[eelw]

What's Vcore at?

 

[BTRY B 529th FA BN]

Sweet

 

[MrTransistorm]

Congratulations!

Actually, many people have reached 5 GHz and beyond. It's the Vcore and temps that are holding most people back.

I also got to 5 GHz on my 2600K. And that was with PLL Overvoltage, Turbo, HT, EIST, C1e, and C3/C6 all enabled. I was even using an offset Vcore instead of fixed. I wouldn't consider it 100% stable, and the voltage is too high for my liking. I have since returned to 4.7 GHz.

 

[Rvenger]

 

[bart1975]

That voltage seems dangerous.

 

[richierich1212]

24/7? What vcore? 1.5v? Going to get degradation.

 

[Arkadrel]

That voltage seems dangerous.

Was thinking the same thing, half a year or less and that cpu is probably dead, if its still running those settings.

1.5v is abit to high.

 

[996GT2]

Was thinking the same thing, half a year or less and that cpu is probably dead, if its still running those settings.

1.5v is abit to high.

24/7? What vcore? 1.5v? Going to get degradation.

Did you guys not read my first post??

 

[RussianSensation]

Thanks for your findings 996GT2.

The most useful takeaways here are that:

1) You don't need a $200-300+ motherboard to max out your SB chip on air cooling.
2) A $30 cooler seems to be sufficient (although you didn't provide any temps, which would have been nice to see at 100% CPU load).

 

[Ado6]

Congrats! Welcome to the 5ghz+ Club :thumbsup:

 

[BTRY B 529th FA BN]

Yeah, I can't wait to stop feeling the need to buy expensive ram for high overclocks like this. The only thing holding me back from buying a 2600K or 2500K is the 23.976 Sandy Bridge issue - although i hear there is some work-around for it but I haven't checked it out yet.

 

[Diogenes2]

How did you do with 32m ? ( time is 7 min 20 sec ).. Temps and volts had already dropped off when I took the screen shot ...

CPU usage is only 12% with prime, so really not much stress going on ..

As far as degradation, Up until a few days ago, I have been folding 24/7 at 4.9 - 5.g & 1.45 -1.48v since late January ,and no problems so far ..
I stopped the CPU folding due to higher summer electric costs ..

http://forums.anandtech.com/showpost.php?p=31156939&postcount=385

I guess time will tell ..

 

[tweakboy]

Launch a intensive game like Crsys 2 and turn off montior and go to sleep.

When you wake up in the morning and its still running and playing then your 100 stable. gl

 

[Patrick Wolf]

Launch a intensive game like Crsys 2 and turn off montior and go to sleep.

When you wake up in the morning and its still running and playing then your 100 stable. gl

Playing GTAIV for 30 min. would be a better test for stability.

 

[Diogenes2]

Launch a intensive game like Crsys 2 and turn off montior and go to sleep.

When you wake up in the morning and its still running and playing then your 100 stable. gl

If I can fold 24/7, and do all the other stuff I need to do, that is all the stable I need...

Why should I care if a game I don't play is stable ..?


If I were not folding, stock would be plenty fast for me, except when I do video encoding, where it can make a significant difference..

 

[aigomorla]

T
2) A $30 cooler seems to be sufficient (although you didn't provide any temps, which would have been nice to see at 100% CPU load).

no you dont.... but the better voltage regs, would allow more efficiency in providing power to your chip, and in hence could allow you to run lower voltages.

Ok... Lemme see if i can explain it to you guys so its easy to understand..

Assume: vcore of 1.3
Assume: Cpu Wattage Load of 150-160W ~ /w Overclocking.
Then... 150/1.3= 115.4 Amps.

Sure the value 1.3v is tiny... its less then a AA battery.
But look at the Amps...

The amount of amps, and clean amps provided is highly dependent on the voltage regs on your board.

So do you really want crappy ones?

Yeah, I can't wait to stop feeling the need to buy expensive ram for high overclocks like this. The only thing holding me back from buying a 2600K or 2500K is the 23.976 Sandy Bridge issue - although i hear there is some work-around for it but I haven't checked it out yet.

yea, i agree with you.... this is a hard habit we need to break...
But higher clocked ram at this point is meaningless because if we play with that QPI, the cpu becomes greatly unstable.

lol my 2600k sits on idle most of the time.. i love it... but i really need a hexie..
And i'll be on a deci soon.

 

[996GT2]

Thanks for your findings 996GT2.

The most useful takeaways here are that:

1) You don't need a $200-300+ motherboard to max out your SB chip on air cooling.
2) A $30 cooler seems to be sufficient (although you didn't provide any temps, which would have been nice to see at 100% CPU load).

Temps at 5.0 GHz and 1.5V would have been quite high with the Hyper 212+ in a closed case, so for this run I pointed a fan at the open case to give it reasonable temps.

Here are what my temps are like in a closed case with the Hyper 212+ and two 1200 RPM fans in push-pull @ 4.5 GHz and 1.344V. Ambient in my apartment is about 76-77C since the AC is off.

Can't wait to get my NH-D14 later this week

 

[Farmer]

no you dont.... but the better voltage regs, would allow more efficiency in providing power to your chip, and in hence could allow you to run lower voltages.

Ok... Lemme see if i can explain it to you guys so its easy to understand..

Assume: vcore of 1.3
Assume: Cpu Wattage Load of 150-160W ~ /w Overclocking.
Then... 150/1.3= 115.4 Amps.

Sure the value 1.3v is tiny... its less then a AA battery.
But look at the Amps...

The amount of amps, and clean amps provided is highly dependent on the voltage regs on your board.

So do you really want crappy ones?



yea, i agree with you.... this is a hard habit we need to break...
But higher clocked ram at this point is meaningless because if we play with that QPI, the cpu becomes greatly unstable.

lol my 2600k sits on idle most of the time.. i love it... but i really need a hexie..
And i'll be on a deci soon.

Using your linear circuit relation and assuming the power, your amperage varies inversely with voltage, i.e., the higher you crank your voltage, the lower your amperage. So, if the point you're trying to make is that high voltage = bad, I don't see how that follows.

You can't assume a fixed power output, you can't model this with just P=IR.

 

[sm625]

How much faster is a 2600k at 5GHz vs a 2500K? Cant be more than 10% right? Surely it cant be worth the extra money.

 

[996GT2]

How much faster is a 2600k at 5GHz vs a 2500K? Cant be more than 10% right? Surely it cant be worth the extra money.

Depends on what you are doing and whether you need hyperthreading. If not, performance is pretty similar between the two at the same clock speeds.

 

[aigomorla]

need the sub 5 second superpi or a 5ghz is moot!

j/k

nice job 996.