i7 2600K @ 5Ghz!

[society misfit]

I’m posting this here instead of the sticky ‘Post Your Overclock’ cause I did not run Prime due to the temperatures I was reaching. So I feel this is not a legit OC. But I did run a stable 30min stress test in AIDA64. Being unfamiliar with Intel, reaching 91+ Celsius scared me enough to back out. However with a better cooler I’m confident that this machine can reach 5GHz+. Apart from setting the DRAM frequencies, everything was done within Asus AI Suite 2 that came bundled with the motherboard.

Previous I could not get past 4.73GHz without BSOD but that was with tweaking the BCLK and multiplier together. Keeping the BCLK to default 100MHz I was able to set the multiplier to 48 no problem on default voltages and run stable with high but respectable temps. Increasing the multiplier any more I could boot into Windows but would BSOD under stress with default voltages.
Tweaking the voltages I was able to set the multiplier to 50 and run what I feel is stable but a HOT 5GHz! Here’s my settings based on a Asus P8P67 Deluxe board.

BCLK = 100MHz
Multiplier = 50X
DRAM was set to 1600MHz @ 1.6V
VCore set 1.37V but throttled up to 1.53V
DIGI+ VRM:
Load Line Calibration = Ultra High
CPU Current Capability = Auto
VRM Frequency = 380Khz
Phase Control = Extreme
Duty Control = Extreme

I’m in the market for a new cooler (advice would be appreciated) so stay tuned!

 

[Everlong]

As I said in the other thread.

Christ.

 

[DaGooN]

nice man. looks a good chip you have. there is some golden chips out there.. 1 is this OC from 1day. lol check this out..

 

[ultimatedesk]

That's pretty amazing. Remember when 4Ghz was the golden "Wow moment".

5Ghz... from a desktop, residential computer, simply amazing.

Has anyone else found that the software that comes with motherboards that allows you to overclock, almost always jumps the voltage too high, and thus increases the temperatures abnormally, compared to manually overclocking?

I'm guessing that it's "safety" on their part, almost guaranteeing a stable OC, but seems a bit overkill. You could probably get some lower temps by manually copying the settings from the software (Into the BIOS) and then slowly creeping up the voltage.

 

[Lordbob75]

That's pretty amazing. Remember when 4Ghz was the golden "Wow moment".

5Ghz... from a desktop, residential computer, simply amazing.

Has anyone else found that the software that comes with motherboards that allows you to overclock, almost always jumps the voltage too high, and thus increases the temperatures abnormally, compared to manually overclocking?

I'm guessing that it's "safety" on their part, almost guaranteeing a stable OC, but seems a bit overkill. You could probably get some lower temps by manually copying the settings from the software (Into the BIOS) and then slowly creeping up the voltage.

This is exactly why overclockers hate overclocking programs...

~Lordbob

 

[society misfit]

The newer Turbo V software that came with the board is a lot better than previous versions. Yes if you click on the “Auto Overclock” button it sets everything very high. But using that function I could not get past 4.7GHz. To reach that magic 5GHz I manually changed the voltages along with various other settings using the Turbo V software. You can make the changes within the software, and if you didn’t crash (or BSOD) right away it allowed you to stress test without having to reboot. I found it very helpful.

I’m currently running at 4.8Ghz 24/7 with very respectable temps. Haven’t had the time to toy around at 5GHz but will do once I pick up a better cooler.

 

[whs]

You better turn this Sandy Bridge back in and wait until they have fixed it.

 

[Mashed1316]

Hmm, how about a Corsair H70? or if you don't want watercooling, you could use a Noctua NH-D14 those would be my #1 choices

 

[Lordbob75]

You better turn this Sandy Bridge back in and wait until they have fixed it.

It is the chip for the SATAII controllers that is bad, not the CPU.

~Lordbob

 

[Mashed1316]

You better turn this Sandy Bridge back in and wait until they have fixed it.

It is the chip for the SATAII controllers that is bad, not the CPU.

~Lordbob

This. i've seen people confused about that issue, as long as you keep your devices away from the SATA II ports, you're fine, plus...

Most end-users don't have more than 4 SATA devices (4 being the SATA III ports on the motherboard) so they wouldn't be that much compromised.

I have 2 HDDs and 2 DVD-Drives (and i only use one, i'm thinking about removing it) i mean, i'd rather have all my HDDs (in case that some users have 4) on the SATA III ports and leave the DVD-Drive(s) on the SATA II, since it's not used often

 

[whs]

You guys are funny. Who wants to pay a lot of money for a defunct CPU - Sata controller, Northbridge, or whatever. Who knows what else is wrong with it. They are only telling us what they want to tell us. And in 32nm technology, there are a lot of opportunities for problems. Let the thing settle down.

 

[Lordbob75]

You better turn this Sandy Bridge back in and wait until they have fixed it.

It is the chip for the SATAII controllers that is bad, not the CPU.

~Lordbob

This. i've seen people confused about that issue, as long as you keep your devices away from the SATA II ports, you're fine, plus...

Most end-users don't have more than 4 SATA devices (4 being the SATA III ports on the motherboard) so they wouldn't be that much compromised.

I have 2 HDDs and 2 DVD-Drives (and i only use one, i'm thinking about removing it) i mean, i'd rather have all my HDDs (in case that some users have 4) on the SATA III ports and leave the DVD-Drive(s) on the SATA II, since it's not used often

To be honest, I am not too concerned about it (until I get a replacement) as the fail rate was so low and over a much longer period of time.
My harddrives should be just fine until I get a replacement motherboard.

~Lordbob

 

[society misfit]

Hmm, how about a Corsair H70? or if you don't want watercooling, you could use a Noctua NH-D14 those would be my #1 choices

I’ve been eyeing the H70. Even shopped around for replacement fans in case the stocks ones are too loud! But both Corsair and Antec will be coming out with newer versions of the “All In One” water cooling very soon. And as for the Noctua NH-D14…I’d be all over that if it weren’t for the tall heatsinks on my memory. It may just fit but I don’t want to take that chance. Plus I like to have a little room to play!

As for the Sandy Bridge fault. I’m not worried at all. So far (from what I’ve read) they have only been able to recreate the failure within a lab environment. With the amount of SB boards that are out there I’ve yet to hear or read about someone coming forth and saying that their SATA II ports actually failed. Besides I’ve got all my “stuff” plugged into the SATA III ports.

 

[Lordbob75]

Hmm, how about a Corsair H70? or if you don't want watercooling, you could use a Noctua NH-D14 those would be my #1 choices

I’ve been eyeing the H70. Even shopped around for replacement fans in case the stocks ones are too loud! But both Corsair and Antec will be coming out with newer versions of the “All In One” water cooling very soon. And as for the Noctua NH-D14…I’d be all over that if it weren’t for the tall heatsinks on my memory. It may just fit but I don’t want to take that chance. Plus I like to have a little room to play!

As for the Sandy Bridge fault. I’m not worried at all. So far (from what I’ve read) they have only been able to recreate the failure within a lab environment. With the amount of SB boards that are out there I’ve yet to hear or read about someone coming forth and saying that their SATA II ports actually failed. Besides I’ve got all my “stuff” plugged into the SATA III ports.

Or better yet a full water cooling setup

Also, see my post in the news thread for motherboard manufacturers response.

~Lordbob

 

[society misfit]

A complete water cooling set up has crossed my mind as well.

BTW, how did you make out overclocking that i5?

 

[Everlong]

You guys are funny. Who wants to pay a lot of money for a defunct CPU - Sata controller, Northbridge, or whatever. Who knows what else is wrong with it. They are only telling us what they want to tell us. And in 32nm technology, there are a lot of opportunities for problems. Let the thing settle down.

Well I know a few people on this forum who bought the mobo before the problem was discovered. And then every vendor/Intel are offering a straight swap for one of the new revision mobos when they're available, and I know places are offering the swap if they're purchased now, so it's not really much of loss, and I doubt anyone will run in to the problem before the new revisions are released. OEMs and Intel found it had a 5% fail rate over extreme usage.

 

[Lordbob75]

A complete water cooling set up has crossed my mind as well.

BTW, how did you make out overclocking that i5?

4.5GHz stable at 1.35V or so.

~Lordbob

 

[society misfit]

A complete water cooling set up has crossed my mind as well.

BTW, how did you make out overclocking that i5?

4.5GHz stable at 1.35V or so.

~Lordbob

Nice. Did you play with the DIGI+ and VRM settings?

 

[Lordbob75]

A complete water cooling set up has crossed my mind as well.

BTW, how did you make out overclocking that i5?

4.5GHz stable at 1.35V or so.

~Lordbob

Nice. Did you play with the DIGI+ and VRM settings?

Not really, I am not sure what those do.

~Lordbob

 

[society misfit]

That makes two of us! But they sound impressive so I’d thought they must do something. Try the settings I posted at the beginning of this thread. But first, go into the BIOS and save your current setting under the profile manager. That way you can go back without any trouble. You can adjust the DIGI+ and VRM in either the BIOS or within the Asus software. Doesn’t make any difference. Once the DIGI+ and VRM are set, save that as a separate profile within the BIOS. This is so you do not have to set those up again if you crash. After that go into the TurboV and try increasing the multiplier 100MHz at a time. After ever 100 MHz increase, do a quick stress test (I used AIDA64) for about 2 or 3min to see where your temps end up at and if it will crash under load. I found that that it would crash seconds within the CPU throttling up to 100%. If you crash it will just revert back to the profile you saved. If you don’t crash the test and you are comfortable with your temperatures, increase again 100MHz. When you crash, take note of the multiplier and either go back 100Mhz and be happy, or increase the VCore by .5 and test again. Once you find a clock you are happy with then you can tear it apart with Prime.

I bet with your VCore set the way it is, by just adjusting the DIGI and VRM you can get another 200MHz out of that i5.