overclocking 5820K Disabling cores for better performance

[TheGamingBarrel]

Just now, Godlygamer23 said:

OC stability is dependent on individual cores. If the cores can't sustain the frequency, OC will not be stable regardless. 

While true, you can achieve a much higher clock speed on less cores for a reason, and  less heat is being produced, therefore increasing stability.

[Godlygamer23]

Just now, TheGamingBarrel said:

While true, you can achieve a much higher clock speed on less cores for a reason, and  less heat is being produced, therefore increasing stability.

Give me evidence that heat automatically correlates to stability. 

[TheGamingBarrel]

1 minute ago, Godlygamer23 said:

Give me evidence that heat automatically correlates to stability. 

When I was running 80C I couldnt get 4.7 Stable at 1.35V, dropped temps to 30C And instantly was doing 5 Stable.

[Godlygamer23]

Just now, TheGamingBarrel said:

When I was running 80C I couldnt get 4.7 Stable at 1.35V, dropped temps to 30C And instantly was doing 5 Stable.

That's not automatic correlation. That's one sample out of thousands, and that would be considered an anecdote. 

[TheGamingBarrel]

2 minutes ago, Godlygamer23 said:

That's not automatic correlation. That's one sample out of thousands, and that would be considered an anecdote. 

Wonder why people use LN2 To achieve the highest clocks? Lower electrical resistance = higher clocks.

[harrynowl]

1 hour ago, Godlygamer23 said:

That's not automatic correlation. That's one sample out of thousands, and that would be considered an anecdote. 

It is the theoretical reason behind people pouring liquid nitrogen on their cpus though. It's about temperature co-efficiencies. So there is some merit to it. 

[Godlygamer23]

Just now, TheGamingBarrel said:

Wonder why people use LN2 To achieve the highest clocks?

Because the die can only handle so much heat. And that's due to the limit of the material being used. They use stupidly high voltages that would probably cause all chips to throttle back if using any kind of mainstream solution.

[TheGamingBarrel]

1 minute ago, Godlygamer23 said:

Because the die can only handle so much heat. And that's due to the limit of the material being used. They use stupidly high voltages that would probably cause all chips to throttle back if using any kind of mainstream solution.

You gain extra clocks at the same volts. I was able to drop voltages and increase clocks, as does EVERYONE on any form of cold.

[Godlygamer23]

1 minute ago, TheGamingBarrel said:

You gain extra clocks at the same volts. I was able to drop voltages and increase clocks, as does EVERYONE on any form of cold.

Everyone? So every single person out there can repeat this?

[TheGamingBarrel]

2 minutes ago, Godlygamer23 said:

Everyone? So every single person out there can repeat this?

Everyone with access to a pot, insulation and LN2, Yes. Even going from air to water can net a big immprovement.

Search up temperature co-efficiencies in circuits

http://www.allaboutcircuits.com/textbook/direct-current/chpt-12/temperature-coefficient-resistance/

[Godlygamer23]

1 minute ago, TheGamingBarrel said:

Everyone with access to a pot, insulation and LN2, Yes. Even going from air to water can net a big immprovement.

I'm not entirely convinced personally as I do not believe heat correlates to stability in this instance. It may occur in biology, but this isn't biology. I won't add my own anecdote because that's all it is. An anecdote. 

[TheGamingBarrel]

3 minutes ago, Godlygamer23 said:

I'm not entirely convinced personally as I do not believe heat correlates to stability in this instance. It may occur in biology, but this isn't biology. I won't add my own anecdote because that's all it is. An anecdote. 

http://hwbot.org Go on their forum and ask people in the chat box or in the sub-zero forum, they will happily teach you. It isnt a anecdote when proven to be true, and not personal.

[CtW]

3 minutes ago, Godlygamer23 said:

I'm not entirely convinced personally as I do not believe heat correlates to stability in this instance. It may occur in biology, but this isn't biology. I won't add my own anecdote because that's all it is. An anecdote. 

 

[harrynowl]

1 hour ago, Godlygamer23 said:

I'm not entirely convinced personally as I do not believe heat correlates to stability in this instance. It may occur in biology, but this isn't biology. I won't add my own anecdote because that's all it is. An anecdote. 

It isn't an anecdote though. It is the entire reason people use sub-ambient cooling  

 

Bit daft to claim something is an anecdote when provided logical scientific reasoning with a source for it and you have nothing to say other than "I'm not convinced". Extreme overclocking is a niche but even as a niche that can't happen without some underlaying scientific principle. People wouldn't spend hundreds on dewars, equipment, probes and such if it isn't something they can actually repeat. 

 

I can test & prove it on my 3570K or this 4770K. 

[TheGamingBarrel]

20 minutes ago, harrynowl said:

It isn't an anecdote though. It is the entire reason people use sub-ambient cooling  

 

Bit daft to claim something is an anecdote when provided logical scientific reasoning with a source for it and you have nothing to say other than "I'm not convinced". Extreme overclocking is a niche but even as a niche that can't happen without some underlaying scientific principle. People wouldn't spend hundreds on dewars, equipment, probes and such if it isn't something they can actually repeat. 

 

I can test & prove it on my 3570K or this 4770K. 

Which 4770K? my one?

[Crossbred]

There's a limit to how far thermals alone can get you in overclocking.

Disabling the cores might make more trouble than it's worth, since the voltage regulators and stuff are made for six cores.

So it could allow for some better thermals, error-preventing, and higher quality processing, but in the end it's a slightly unstable 6700k with 8 ram slots imo.

Just being a devil's advocate, really good thinking on your part though.

[harrynowl]

1 hour ago, TheGamingBarrel said:

Which 4770K? my one?

yes that one

[TheGamingBarrel]

8 minutes ago, Crossbred said:

There's a limit to how far thermals alone can get you in overclocking.

Disabling the cores might make more trouble than it's worth, since the voltage regulators and stuff are made for six cores.

So it could allow for some better thermals, error-preventing, and higher quality processing, but in the end it's a slightly unstable 6700k with 8 ram slots imo.

Just being a devil's advocate, really good thinking on your part though.

Nope, stable 4790K. haswell-e.

[Crossbred]

Just now, TheGamingBarrel said:

Nope, stable 4790K. haswell-e.

ddr4

[done12many2]

5 hours ago, jwakeford said:

Hi, 

I am planing on a new build for this summer, I will be using my PC almost exclusively for gaming. 

At first I thought about going for a 6700K but as I got more informed about the grand scheme of things in the tech world I got to the conclusion that the right path was the 5820K : 

- Direct X12, Vulcan and future game optimizations all point towards a better spreading of the work load throughout the CPU cores, giving multi-core performance a go against single core performance. 
- 5820K when overclocked almost matches an overclocked 6700K. 
- The 5820K has a better value in Price/performance overall. 
- And finally, the point I would like to confirm :  
          You can disable two cores on the 5820K, preferably cherry picking the two worse performing core, and therefore increasing the overall remaining silicon quality. 
           This makes the 4 core 5820K more stable, and easier to cool down as less cores produce less heat. 
All this should result in a quite significant advantage for the 5820K over the 6700K when overclocked. 

What do you think ? 
Is it possible ? Advisable to do so? 

Thanks for your input. 

OP, there sure is a lot speculating at best going on in this thread.  With 6 cores and higher, your biggest battle and limiting factor in overclocking is going to be heat.  It is very common for those with 6 or more cores to turn off unneeded cores in order to to achieve higher clock rates.    

 

No, turning off cores wont overcome the limits set by the quality of your chip, but adding voltage while reducing/maintaining heat does within reason, therefore your overclock will in most cases increase at a lower core count.

 

My particular chip does well with 8 cores active at 4.9 GHz, however anything above that requires dropping cores.  Not so much due to heat because my cooling setup tames that well, but it just runs faster on 4 cores with more voltage than it does 8.

 

Also, generally speaking, most people with a 5820k, 5930k, or 5960x don't disable cores because most people with these setups are using their rigs for purposes other than benchmarking wars and the such.  On the occasions that I beat out a quad core, it's usually because I was only running 4 cores at the time.  That's my experience and I believe it to be directly relevant to what you're asking.

[done12many2]

5 hours ago, Ronnie76 said:

You can do individual core overclocking

Disabling 2 cores is literally the stupidest thing you could EVER do 

Remember, if every core is producing 60c the core temp wont go above 60c, 

6 cores at 60c each is an average of 60c

4 cores at 60c is an average of 60c

Disabling cores really will not help you.

Yeah, not so much.  You're actually way off on your guess.  The same thing will happen with your 4790k if you went from 4 to 2.  Average and peak temp will drop.  This is a fact.  

 

Making bold and profound statements like "is literally the stupidest thing you could EVER do" doesn't make anyone believe your guess on what would happen any more believable.

 

The OP is absolutely correct about the fact that reducing cores lowers overall temps.  A 5960x at 4 cores requires less voltage, produces less heat and in my case definitely overclocks higher than it would at 8 cores.  

 

If I want to run a benchmark that benefits from single threaded performance, I turn off some cores and crank up the multiplier.  If I am going to run Cinebench or some other multi threaded benchmark, I use all cores at a lower clock rate.  Trust me, I'd love to keep it at the same clock rate as I would with 4 cores, but it's just not happening.

 

With all of this said, I ran a couple of really quick AIDA64 CPU stresses to show an apples to apples comparison.  Note that in addition to the fact that average temps (graph) took a notable drop with 4 cores, the VRM was running substantially cooler.  I would say that water temp was a negligible difference in favor of 4 cores.

 

As the OP was suggesting, 4 cores would run cooler than 6, would require less voltage at the same given frequency, and could potentially overclock better when single core/thread performance matters.

 

I don't see anywhere in his post where he suggested that 4 cores at a higher frequency was more powerful than 6 at a lower.  I just think he understands that there are benefits to being able to curtail your overclock and core use to what you need at that moment.

 

[Curufinwe_wins]

5 hours ago, Godlygamer23 said:

That's not automatic correlation. That's one sample out of thousands, and that would be considered an anecdote. 

This is how silicon works.

 

This is how QM (aka reality) works. See the Maxwell-Boltzmann equation.

 

As temperature increases, average energy increases, thus increasing electron leakage from transistors, thus reducing stability.

 

NOTE: Massive over-simplification, but it's a correct first order approximation. Without detailed knowledge of electronics and QM you won't be able to answer the question in more detail.

[done12many2]

To further demonstrate that the OP isn't making shit up, here is the same processor at 5 GHz and even more voltage.  As you see in the shots from the 4 cores running at the the same voltage and speed as the 8 cores and the shot of the 4 cores running with more voltage and speed then the 8 cores, temperatures are far more manageable in 4 core compared to 8.  It's a fact.  I can barely boot into Windows at any voltage with 8 cores at 5 GHz, but I can use 4 cores at 5 GHz or higher with stability.

 

[Curufinwe_wins]

19 minutes ago, done12many2 said:

To further demonstrate that the OP isn't making shit up, here is the same processor at 5 GHz and even more voltage.  As you see in the shots from the 4 cores running at the the same voltage and speed as the 8 cores and the shot of the 4 cores running with more voltage and speed then the 8 cores, temperatures are far more manageable in 4 core compared to 8.  It's a fact.  I can barely boot into Windows at any voltage with 8 cores at 5 GHz, but I can use 4 cores at 5 GHz or higher with stability.

I hates you...

 

I can run 4.9 at 1.45 V on my 5820k all 6 cores, but my IMC is so bad I can't even do DDR4 2666 with any kit I've tried. I also can't push even 1 MHz out of spec on bus clocks.

[jwakeford]

1 hour ago, done12many2 said:

Yeah, not so much.  You're actually way off on your guess.  The same thing will happen with your 4790k if you went from 4 to 2.  Average and peak temp will drop.  This is a fact.  

 

Making bold and profound statements like "is literally the stupidest thing you could EVER do" doesn't make anyone believe your guess on what would happen any more believable.

 

The OP is absolutely correct about the fact that reducing cores lowers overall temps.  A 5960x at 4 cores requires less voltage, produces less heat and in my case definitely overclocks higher than it would at 8 cores.  

 

If I want to run a benchmark that benefits from single threaded performance, I turn off some cores and crank up the multiplier.  If I am going to run Cinebench or some other multi threaded benchmark, I use all cores at a lower clock rate.  Trust me, I'd love to keep it at the same clock rate as I would with 4 cores, but it's just not happening.

 

With all of this said, I ran a couple of really quick AIDA64 CPU stresses to show an apples to apples comparison.  Note that in addition to the fact that average temps (graph) took a notable drop with 4 cores, the VRM was running substantially cooler.  I would say that water temp was a negligible difference in favor of 4 cores.

 

As the OP was suggesting, 4 cores would run cooler than 6, would require less voltage at the same given frequency, and could potentially overclock better when single core/thread performance matters.

 

I don't see anywhere in his post where he suggested that 4 cores at a higher frequency was more powerful than 6 at a lower.  I just think he understands that there are benefits to being able to curtail your overclock and core use to what you need at that moment.

 

Thank you very much for your answer, you really clarified my doubts about the matter. 
I also really appreciate you taking your time to bring some actual proofs to settle the matter. 

In the end, I will go with a 5820K !  
-> Modularity 
-> Future proofing with multicore hungry games
-> Value just better paying the same as 6700k but +2 cores and double the threads all while still being comparably overclocked.