SB Overclocking results?

5Ghz @ 1.45v

5ghz.png


4.8Ghz @ 1.35v which are my everyday settings

48ghz.png


1253 BIOS Settings:

AI Tweaker:
AI OC- Auto
Turbo Raito- Per Core
Internal PLL- Enabled
Mem Freq- Per your RAM specs
Load Line- Ultra High
VRM Freq- Manual
VRM Fixed- 350
Phase- Extreme
Duty- Extreme
CPU Volt- Manual, 4.8Ghz @ 1.35v, 5.0Ghz @ 1.45v
DRAM Volt- Per your RAM specs

Advanced --> CPU Config
CPU Ratio- Auto
Active Cores- All
Limit CPUID- Disabled
Everything else- Enabled
 
Last edited:
Pretzel said:
5Ghz @ 1.45v

what are your temps under full load while testing. When you took that screen shot it appears that the burntest was completed so your chip would have started to or been cooled down by that time.

Also I was having BSOD at 5ghz at 1.4volt. I am a little nervous to run higher than 1.4 because of fluctuation.

Also is there a specific reason for going 100.3 bclk?

And lastly at 4.9ghz I can run burn test and prime 95 all freaking day long but I get BSOD when I do something simple like load firefox. it doesnt make any sense. Besides passing burn test etc... have you experienced any BSOD?
Click to expand...
 
tangoseal said:
what are your temps under full load while testing. When you took that screen shot it appears that the burntest was completed so your chip would have started to or been cooled down by that time.
Click to expand...

I had speed step enabled so I turned on another burn test in the background to display the correct Ghz in CPU-Z once the burn test was completed. It shows my max temps in Real Temp in those screen shots. 5.0Ghz 79C and 4.8Ghz 67C. I too would have a BSOD at 5Ghz at 1.4v and thus needed to bump to 1.45v to be stable. The BCLK is set at 100 in the bios, you will see slight variations displayed in CPU-Z. I haven't messed around too much above 4.8Ghz other than to run synthetic tests. I find that 4.8Ghz is the sweet spot to be able to keep stable at low voltage. 4.9 and up calls for much higher voltages to stay stable.
 
Pretzel said:
I had speed step enabled so I turned on another burn test in the background to display the correct Ghz in CPU-Z once the burn test was completed. It shows my max temps in Real Temp in those screen shots. 5.0Ghz 79C and 4.8Ghz 67C. I too would have a BSOD at 5Ghz at 1.4v and thus needed to bump to 1.45v to be stable. The BCLK is set at 100 in the bios, you will see slight variations displayed in CPU-Z. I haven't messed around too much above 4.8Ghz other than to run synthetic tests. I find that 4.8Ghz is the sweet spot to be able to keep at low voltage. 4.9 and up calls for higher voltages to stay stable.
Click to expand...

What v are you running at 4.8? I am just trying to assess my own settings. Thanks for the prompt reply.
 
Pretzel said:
5Ghz @ 1.45v

5ghz.png


4.8Ghz @ 1.35v which are my everyday settings

48ghz.png


1253 BIOS Settings:

AI Tweaker:
AI OC- Auto
Turbo Raito- Per Core
Internal PLL- Enabled
Mem Freq- Per your RAM specs
Load Line- Ultra High
VRM Freq- Manual
VRM Fixed- 350
Phase- Extreme
Duty- Extreme
CPU Volt- Manual, 4.8Ghz @ 1.35v, 5.0Ghz @ 1.45v
DRAM Volt- Per your RAM specs

Advanced --> CPU Config
CPU Ratio- Auto
Active Cores- All
Limit CPUID- Disabled
Everything else- Enabled
Click to expand...

Hey Pretzel, your not running the IBT right and could cook your CPU in a high pressure game.

You setings should be at lease:
Stress Level >>> Very High
Threads>>>>> Manually set to 8

You won't need to run but 3 times to get a good pass or fail. Be ready to stop the burn as I have a feeling your temps are going to go in the 90c range.
You have a good chip but they are not that good on Air. ;) The 4.8GHz clock should be safe cause the VCore is low.
 
Is it me or is my system weird, because I consistently get higher temps with Prime95's Small FFTs test versus IBT, even with the newest version of linpack from Intel and maximum stress level with 8 threads?
 
Armpit said:
Is it me or is my system weird, because I consistently get higher temps with Prime95's Small FFTs test versus IBT, even with the newest version of linpack from Intel and maximum stress level with 8 threads?
Click to expand...

Prime95 is definitely harder on these chips.
 
Armpit said:
Is it me or is my system weird, because I consistently get higher temps with Prime95's Small FFTs test versus IBT, even with the newest version of linpack from Intel and maximum stress level with 8 threads?
Click to expand...
Try IBT with 4 threads, that's usually more stressful for HT Sandy Bridges.
 
The newest version of IBT using threads set to automatic instead of 8 will produce higher temps than IBT with 8 threads or Prime 95.
 
SilverSliver said:
The newest version of IBT using threads set to automatic instead of 8 will produce higher temps than IBT with 8 threads or Prime 95.
Click to expand...

I tried putting the newer version of Linpack in, and I still get higher loads (more consistent) and higher temps with Prime95.
 
People need to run Prime95 for at least 1 hr. IBT doesn't really prove stability with 2500K/2600K since you can run IBT for hours and fail Prime95 in minutes.
 
Myzhi said:
People need to run Prime95 for at least 1 hr. IBT doesn't really prove stability with 2500K/2600K since you can run IBT for hours and fail Prime95 in minutes.
Click to expand...

I thought that to. I let the Prime95 do the work overnight and it passed for 12 hours no problem. 15 minutes later the OS froze like some crazy bitch. I don't believe in high stable OC at all.
 
I always use prime95 to see if my system is stable. I always felt p95 was the golden standard for that.
 
Some findings / advice I want to share.

While trying countless overclock settings, I've been having a lot of crashes going into Windows, mostly at the windows logo. Straight up lock then BSOD,

But, on one of those attempts, I moved and unplugged an external hard-drive. And that particular overclock ... worked. I noticed this right away. So I tried it plugged in again, with the same settings and .... hard lock at the logo again.

I know there are lots of you out there having this same issue. I've read your posts.

Definitely unplug / turn off any external device before trying your overclock settings. Printers, USB anything.

See if this helps.

This is with the Asus P8P67 Deluxe
 
Last edited:
SixFootDuo said:
Some advice I want to share.

While trying countless overclock settings, I've been having a lot of crashes going into Windows, mostly at the windows logo. Straight up lock then BSOD,

But, on one of those attempts, I moved and unplugged an external hard-drive. And that particular overclock ... worked. I noticed this right away. So I tried it plugged in again, with the same settings and .... hard lock at the logo again.

I know there are lots of you out there having this same issue. I've read your posts.

Definitely unplug / turn off any external device before trying your overclock settings. Printers, USB anything.

See if this helps.
Click to expand...

An external hard drive I can understand but printer? So no overclockers can ever print anything???
 
I wouldn't focus so much on the printer. What I am saying is, "any" usb device could be causing a problem with your overclock attempts. I am sure most people were able to get that from my post.
 
Some devices get very funny with usb3 ports, like my mouse for instance. Try without using any usb3 ports,
 
SixFootDuo said:
I wouldn't focus so much on the printer. What I am saying is, "any" usb device could be causing a problem with your overclock attempts. I am sure most people were able to get that from my post.
Click to expand...

I don't know where you are getting that. I've never heard of any USB devices messing with overclocks. I have a ton of USB ports plugged in and absolutely no issues.
 
Don't know where I'm getting ....... whaaaa ?

I put it in my post, I got it from my "own" experience .... that I could duplicate and can still duplicate. Are you one of those people that hear what people say the first time but still say "what?"
 
hmz said:
I thought that to. I let the Prime95 do the work overnight and it passed for 12 hours no problem. 15 minutes later the OS froze like some crazy bitch. I don't believe in high stable OC at all.
Click to expand...

I don't even bother with prime on the 2600k. The temps are too high. If you can pass Deep Rybka 4 or another multi-threaded chess engine without crashes, then you should be able to pass anything or any game, that isn't a folding app. Rybka gets the cpu's to 75C at 5 ghz and 1.4v (1.432 in cpuz); Prime gets them to 88C+. I can't get the temps lower, though...(tried both a Ven-X and a True...and I'm NOT sticking my 190 cfm Delta fan in there without a fan controller. Doubt it would make much difference anyway...

That 13C difference between real world application (Rybka/chessbase) and Benchmark stress tester is important.
 
The problem I see with these "stable-for-what-I-do" overclocks are that you then have to worry about pushing your computer too hard.

I like the peace of mind that stress tests provide.

I never have to say to myself "Wait, I should not do this because it's too CPU intensive."
 
Yeah, to me, it's not stable unless it can pass Prime95 for 8+ hours anyways... None of this run a test for 5 minutes nonsense. I've see way too many overclocks fail that way.
 
SixFootDuo said:
I wouldn't focus so much on the printer. What I am saying is, "any" usb device could be causing a problem with your overclock attempts. I am sure most people were able to get that from my post.
Click to expand...

On both my LGA775 system and my new 1155 system, if I leave my external HDD plugged in when the system restarts, it can cause a frozen or slow POST.
 
I've never had that problem before, ever. But sure enough with my new Asus P9P67 Deluxe board, if I have an external drive plugged in, I lock up at the Windows 7 logo.

Since I've flashed to the new 1253 bios and unplugged my external drive, I've been able to O/C to 4.8Ghz at 1866 Mem. Seems stable so far from the handful of stress tests I've done.
 
Database said:
Yeah, to me, it's not stable unless it can pass Prime95 for 8+ hours anyways... None of this run a test for 5 minutes nonsense. I've see way too many overclocks fail that way.
Click to expand...

That's why I run something like a game, or my Rybka chess program :) If that passes then I'm good to go. And it's alot more fun to play various games than run Prime. But whatever floats your boat. Just please don't impose your values on me, unless you're willing to pay for my hardware.

Why should I run something that will degrade my CPU at 90 degrees C, if I can keep my temps at 75C or lower?

I've degraded enough chips from stress testing to know better.
 
In the past, I would always loop 3Dmark 200x and prime95 at the same time, overnight (at least 8 hours, usually more) before I would consider an over-clock acceptable.
 
Falkentyne said:
That's why I run something like a game, or my Rybka chess program :) If that passes then I'm good to go. And it's alot more fun to play various games than run Prime. But whatever floats your boat. Just please don't impose your values on me, unless you're willing to pay for my hardware.

Why should I run something that will degrade my CPU at 90 degrees C, if I can keep my temps at 75C or lower?

I've degraded enough chips from stress testing to know better.
Click to expand...

And don't expect me to help when you make your threads stating random BSOD, restarts, and lock ups.
 
Database said:
And don't expect me to help when you make your threads stating random BSOD, restarts, and lock ups.
Click to expand...

Why would he need to start a thread? If he is playing games and gets a bluescreen, up the voltage and check for stability. Stress testing software is a great tool, but it's just that, a tool. If he's stable with the temperatures he is comfortable with in everyday usage, I don't see what the problem is. If he was a 24/7 folder or a heavy transcoder, it'd be a different story.
 
SilverSliver said:
Why would he need to start a thread? If he is playing games and gets a bluescreen, up the voltage and check for stability. Stress testing software is a great tool, but it's just that, a tool. If he's stable with the temperatures he is comfortable with in everyday usage, I don't see what the problem is. If he was a 24/7 folder or a heavy transcoder, it'd be a different story.
Click to expand...

Eventually he will crash and probably come here posting about it blaming something other than his "stable for me overclock". If you cant pass 8- 10 hours of prime and linx then you just aren't 100% stable regardless of what you do with it.
 
dnottis said:
Eventually he will crash and probably come here posting about it blaming something other than his "stable for me overclock". If you cant pass 8- 10 hours of prime and linx then you just aren't 100% stable regardless of what you do with it.
Click to expand...

I've never run 10 hours in Linpack and I don't think I would recommend it to anyone, regardless of what overclock you are running. I did an hour of Prime and 10 passes of very high load Linpack with my 24/7 overclock. I've not had one bluescreen in days since I found the stable voltage. Transcoding, lots of games, folding, you name it. You do not need 10 hours of Linpack to prove or disprove stability.Furthermore,you are making a lot of assumptions about a guy you don't even know. If he is crashing, he will just up the volts or lower the overclock. It's not like it's a quantum physics problem. If you are stable at 10 runs of Linpack and you bluescreen anyhow, it's probably only a difference of .005v or something ridiculous that will make you permanently stable. Linpack is designed to prove stability or instability in a shorter amount of time. That's the goal of any stability program. Instead of modifying values and doing normal usage and waiting for a bluescreen.

I mean, if you are going through the trouble of proving 10 hour Linpack stability, you must have some pretty extreme usage or environmental requirements. Obviously, 10 hours of Linpack alone is not enough. Try 10 hours of Linpack with a room temperature of above 90F. Then try 10 hours of Linpack with a room temperature of below 20F. Then try Linpack in each case with a humidity level of 75% at both 90F and 20F. Then try Linpack 10 hours in a Vacuum. Then try Linpack 10 hours while your computer is adjusted to 8 orientations. Then try Linpack for 10 hours while using the computer as a commercial hotdog roaster. Then try Linpack for 10 hours while it rides on a surfboard pulled by a shark with a water proof UPC strapped around it's neck.

Then, and only then, is your overclock stable.
 
Last edited:
You must log in or register to reply here.
Back
Top