SB Overclocking results?

[exdriver]

I believe that when we put over the specs. Vcore on cpu, the capacity of cooling should increase too. Just to slow it eventual degradation

I firmly believe you're right. Eventual can become quick if you have enough heat.

Like what you're running there... I doubt your chip would survive long at 1.6V at 80C.
With cooling like yours, well... over VID max might not be a problem for quite a while.

 

[centvalny]

I firmly believe you're right. Eventual can become quick if you have enough heat.

Like what you're running there... I doubt your chip would survive long at 1.6V at 80C.
With cooling like yours, well... over VID max might not be a problem for quite a while.

QFT. Again, I never run it with that settings for 24/7

Just for benching purposes and then go back to all default settings

 

[theone1989]

I find that with my i5-2500k it hits 4.4ghz stable with almost no effort. I did have to tweak to get 4.5ghz stable. currently at 1.34v in BIOS or max 1.352 underload in CPU Z with VCCIO @1.11v also.

But to get it to 4.6ghz and not crash less than 30mins into Prime 95, I've had to set vcore to 1.36 in BIOS with max draw of 1.368 in Prime Blend, and probably going to be needing 1.376v in Small FFT. I've also had to set VCCIO to 1.15v, as well as disabling C1E support (which means multi is at full 4.6ghz all the time).

So it seems to me going over 4.5ghz isn't too practical in 24/7 day to day usage. Especially given how much I'm having to tweak and drive the CPU to get that extra 100mhz. God knows what I'll have to do to make it hit 5.0ghz..

 

[exdriver]

But to get it to 4.6ghz and not crash less than 30mins into Prime 95, I've had to set vcore to 1.36 in BIOS with max draw of 1.368 in Prime Blend, and probably going to be needing 1.376v in Small FFT. I've also had to set VCCIO to 1.15v, as well as disabling C1E support (which means multi is at full 4.6ghz all the time).

So it seems to me going over 4.5ghz isn't too practical in 24/7 day to day usage. Especially given how much I'm having to tweak and drive the CPU to get that extra 100mhz. God knows what I'll have to do to make it hit 5.0ghz..

Pretty much the consensus on the sweet spot for these chips and that's what the [H] article on it said. Generally speaking I think they know what they're talking about

It could be practical, with the right cooling, practical being relative. Nothing really practical about a computer that costs more than say a few hundred dollars I guess. I don't think you'd cause any real harm to the chip if you keep your temps in line if you are near the maximums given by Intel. FWIW they do say that each chip is different.

Watch the load temps on your motherboard, though. I have a pretty spacious case and a ton of airflow and could still get the board fairly warm.

 

[glen]

When compared to famous overclockers like the D 805 or the Q9550, these processors aren't really producing reliable stable overclocks to be impressed with. Meaning stable overclock with stock voltage and all cpu power saving features on. If you are lucky, 4.4ghz. But even with that folks are seeing stability issues. And what is very concerning is the quick degredation being reported with high , stable overclocks requiring exceptional voltage. I would shoot for 4.1 with everything left stock until the dust settles. Way too many unexplained stability issues on the web. Remember when you bump up that multiplier the warranty is gone, and these chips don't just die they gradually get slower as a failure mode. I even believe the C states are there to lengthen lifespan rather than save power. My hunch is that the 32nm archeticture is not as robust as others have been?

 

[Blahman]

When compared to famous overclockers like the D 805 or the Q9550, these processors aren't really producing reliable stable overclocks to be impressed with. Meaning stable overclock with stock voltage and all cpu power saving features on. If you are lucky, 4.4ghz. But even with that folks are seeing stability issues. And what is very concerning is the quick degredation being reported with high , stable overclocks requiring exceptional voltage. I would shoot for 4.1 with everything left stock until the dust settles. Way too many unexplained stability issues on the web. Remember when you bump up that multiplier the warranty is gone, and these chips don't just die they gradually get slower as a failure mode. I even believe the C states are there to lengthen lifespan rather than save power. My hunch is that the 35nm archeticture is not as robust as others have been?

I do believe you may be speaking from your backside, sir.

 

[theone1989]

Pretty much the consensus on the sweet spot for these chips and that's what the [H] article on it said. Generally speaking I think they know what they're talking about

It could be practical, with the right cooling, practical being relative. Nothing really practical about a computer that costs more than say a few hundred dollars I guess. I don't think you'd cause any real harm to the chip if you keep your temps in line if you are near the maximums given by Intel. FWIW they do say that each chip is different.

Watch the load temps on your motherboard, though. I have a pretty spacious case and a ton of airflow and could still get the board fairly warm.

Well for one thing, I'm beginning to suspect the 2600k must have been speed binned because a lot of high overclockers have the 2600k instead of my 2500k. They seem to report 100mhz or 200mhz higher overclock at same voltage.

When I was running my chip at 4.6ghz at 1.368v and VCCIO of 1.15, the temp on the hottest core never exceeded 60 maximum and it was usually around 58/59, the other cores hovering around 52 and 55. This is on air with my Zalman CPNS9900MAX

To be honest, given how many people seem very comfortable running it under load hitting the 70s and 80s running it at 1.4v etc, I could probably push mine further if I set vcore even higher then starting upping the System Agent voltage, then the PLL voltage as well.

But then I think, hmm do I really need that extra multiplier or 2 of performance. Oh, I did notice something, it seems that an extra 100mhz will give you 2 extra gigaflops of speed on Intel Burn Test.

 

[glen]

I do believe you may be speaking from your backside, sir.

Hope you are right, I'm just not cut out to be an early adopter like I used to be and we really haven't established what is safe on these CPU's yet. Only a big warning from C-N concerning extended benchmarking. I do know why there is no excplicit max voltage listed by Intel for these CPU's.

 

[exdriver]

Many people are running overclocks that exceed the limits for voltage given for cpu and ram. If you do that, and you have problems, take a post it note and write they told you so on it. If you are worried about your warranty then overclocking is a pretty moot point to begin with and it is very likely that your chip will outlive your favorite dog or cat with the benefit of never having to go to the vet. If you're going to stay within the limits suggested 4.1 is a little low, I think you can run 4.3~4.4 like has been said without issue.

The stock cooler's expected performance at full tdp is around 72C if I read the paper right. I would think as long as you're under this you're probably ok and again with decent air cooling even at full tdp you should be around this mark with an overclock like this. Day to day usage your temps are going to be less.

And no, I agree, you don't need the extra multi or two. Not for the price of volts and temps. When you do need it, its probably time to upgrade again anyway. I'd think they are speed binned too, especially given that they slapped HT on one and not the other.

If you want to be all practical about it

 

[Rossi~]

*bump*

 

[MaZa]

Hope you are right, I'm just not cut out to be an early adopter like I used to be and we really haven't established what is safe on these CPU's yet. Only a big warning from C-N concerning extended benchmarking. I do know why there is no excplicit max voltage listed by Intel for these CPU's.

Yes there is, asked from Intel.

http://forums.overclockers.co.uk/showpost.php?p=18190523&postcount=103
branches here http://forums.overclockers.co.uk/showthread.php?t=18227651

Whole thread is interesting to read too.

Intel strictly recommends not to go over 1.38v on Vcore. 1.4 and above is asking for trouble. Some can take higher (though for how long?) but not all. This is after all 32nm CPU we are dealing with. And especially BCLK overclocking is now known to be a CPU killer, so leave it at 100mhz.