4790K at 1.22 V clock speeds

[wiebevb]

Hi,

 

I am running my 4790K at 1.22 volts and I overclocked it to 4.6 GHz on all cores. I am using this overclock for 4 days now and it seems stable. Also, i runned AIDA64 stress tests.

 

Is it possible that the 4790K runs at 4.6 GHz on 1.22 V? I saw other people on the form that were using much higher core voltages (like 1.28).

 

[Queek]

Are you using manual or adaptive voltage mode?

It's possible that you won the silicon lottery, yes.  

[vembutech]

You are lucky person. 

[-rascal-]

It certainly is possible -- remember, each CPU is different.

Not a i7-4790K, but my new i5-4690K runs at 4.6GHz with 1.2 ~ 1.21V....but seems to need a little over 1.25V for 4.7GHz.

[incarnate]

yes it is possible. my i5 hit 4.5 at 1.2v which is a bit better than average and 4.4 at 1.165. you seem to have a very nice overclocker there. maybe some day i'll try pushing 4.6 or 4.7 just to see if i can hit it before 1.3v

[incarnate]

also, what do you have your LLC set to, if you left it on auto it might be using higher settings which brings your voltage above your specified voltage under load

[Gargantua]

i was running mine at 4.6ghz on 1.17v. its absolutly possible .coz i don't need overclock so i did downclocked it . and now its on stock but 1.13v.

 

[wiebevb]

Are you using manual or adaptive voltage mode?

It's possible that you won the silicon lottery, yes.  

 

I am using manual voltage mode.

 

 

also, what do you have your LLC set to, if you left it on auto it might be using higher settings which brings your voltage above your specified voltage under load

 

I do not know what LLC is, but here are some screens in CPU-Z and Intel XTU. Maybe you can look it up? My motherboard is a MSI Z97 Gaming 5

 

 

 

 

yes it is possible. my i5 hit 4.5 at 1.2v which is a bit better than average and 4.4 at 1.165. you seem to have a very nice overclocker there. maybe some day i'll try pushing 4.6 or 4.7 just to see if i can hit it before 1.3v

 

I AIDA64, core temps reached around 72 degress celcius in a room temperature of 21 degress celcius. In the evening, they can hit 80 degress, when it's warm outside. So maybe I do a higher overclock in the winter. I think I am using this CPU for a long time, so my comfort zone is around 80 degress.

 

My cooler is a Noctua NH-D14

[incarnate]

I am using manual voltage mode.

I do not know what LLC is, but here are some screens in CPU-Z and Intel XTU. Maybe you can look it up? My motherboard is a MSI Z97 Gaming 5

I AIDA64, core temps reached around 72 degress celcius in a room temperature of 21 degress celcius. In the evening, they can hit 80 degress, when it's warm outside. So maybe I do a higher overclock in the winter. I think I am using this CPU for a long time, so my comfort zone is around 80 degress.

My cooler is a Noctua NH-D14

I highly recommend you read this article

http://linustechtips.com/main/topic/24019-load-line-calibration-why-overclockers-should-care/

It explains what llc is. At 1.22v you should be ok but basically llc adds voltage when under load. If you haven't touched the setting your cpu is likely using more voltage than you think

[Gargantua]

I am using manual voltage mode.

 

 

 

I do not know what LLC is, but here are some screens in CPU-Z and Intel XTU. Maybe you can look it up? My motherboard is a MSI Z97 Gaming 5

 

 

 

 

 

I AIDA64, core temps reached around 72 degress celcius in a room temperature of 21 degress celcius. In the evening, they can hit 80 degress, when it's warm outside. So maybe I do a higher overclock in the winter. I think I am using this CPU for a long time, so my comfort zone is around 80 degress.

 

My cooler is a Noctua NH-D14

D.C cpus on Z97 are mad dogs in voltage (underload) regard . i don't know what is wrong with it , is it bios or cpu code it self but you have to set voltage manually if you want low temps and CPU will work fine too but on stock (Auto) voltage your CPU will easily hit 1.3+ voltage under load on stock .D-14 is an awesome cooler by the way. don't run on stock ( auto voltage). 72c underload on 4.6 at 1.22V is normal .

[wiebevb]

D.C cpus on Z97 are mad dogs in voltage (underload) regard . i don't know what is wrong with it , is it bios or cpu code it self but you have to set voltage manually if you want low temps and CPU will work fine too but on stock (Auto) voltage your CPU will easily hit 1.3+ voltage under load on stock .D-14 is an awesome cooler by the way. don't run on stock ( auto voltage). 72c underload on 4.6 at 1.22V is normal .

I am not running on auto voltage, but manual, so I should be fine I think.

 

I highly recommend you read this article

http://linustechtips.com/main/topic/24019-load-line-calibration-why-overclockers-should-care/

It explains what llc is. At 1.22v you should be ok but basically llc adds voltage when under load. If you haven't touched the setting your cpu is likely using more voltage than you think

Ok, I checked the voltage readings in AIDA64 and CPU-Z when running a short stress test. They both never reach above 1.232 V (they give the same voltages, see pic)

On idle, the voltage is 1.216. That extra voltage can be applied by LLC.

I think 1.23 V is still great for a 4.6 GHz overclock.

 

[i_build_nanosuits]

Is it possible that the 4790K runs at 4.6 GHz on 1.22 V? I saw other people on the form that were using much higher core voltages (like 1.28).

yes it's absolutely possible, and 1.28v is not ''much higher voltages'' it's ''very slightly higher voltage''...

[Queek]

I highly recommend you read this article

http://linustechtips.com/main/topic/24019-load-line-calibration-why-overclockers-should-care/

It explains what llc is. At 1.22v you should be ok but basically llc adds voltage when under load. If you haven't touched the setting your cpu is likely using more voltage than you think

It gives me a warm fuzzy feeling knowing that people still remember that article 

Unfortunately LLC has become irrelevant, as Haswell and Devil's Canyon voltage regulation happens on the chip itself.  LLC now affect Vin only (the voltage supplied to the chip), and I have found that different LLC settings no longer affect stability.  

 

 

I am using manual voltage mode.

It appear that you have a very nice chip! Congratulations!

You should see how high you can get it, for science.  That way those of us with only mediocre chips can live vicariously through you.  

[wiebevb]

It gives me a warm fuzzy feeling knowing that people still remember that article 

Unfortunately LLC has become irrelevant, as Haswell and Devil's Canyon voltage regulation happens on the chip itself.  LLC now affect Vin only (the voltage supplied to the chip), and I have found that different LLC settings no longer affect stability.  

 

 

It appear that you have a very nice chip! Congratulations!

You should see how high you can get it, for science.  That way those of us with only mediocre chips can live vicariously through you.  

 

That's great to hear

 

Maybe i try to overclock it to 4.7 GHz on the same voltage level, test if it is stable . I should probably add more voltage, but I can try.

Does higher clock speeds give a higher temperature, or by upping the the voltage level? And for a long period, how is my CPU lifetime affected?

[Queek]

Does higher clock speeds give a higher temperature, or by upping the the voltage level? And for a long period, how is my CPU lifetime affected?

Both will give higher temperature.  This is why it's so difficult to overclock to the maximum of a system, you have to balance temperature against performance.  

 

There are varying opinions on how a cpu's lifetime will be affected by overclocking, and no definitive answer that I've heard.  One thing you should do after you finish validating your overclock is to switch it over to adaptive voltage mode.  In adaptive mode, your cpu will only go to that maximum voltage and frequency when it needs it, so most of the time it will be at low temperatures and voltage, which will certainly extend your cpu's lifespan.  

If you have any questions on adaptive mode, mention me in a post and I can help you out.  Adaptive setup is really easy, but not very intuitive.  

[wiebevb]

Both will give higher temperature.  This is why it's so difficult to overclock to the maximum of a system, you have to balance temperature against performance.  

 

There are varying opinions on how a cpu's lifetime will be affected by overclocking, and no definitive answer that I've heard.  One thing you should do after you finish validating your overclock is to switch it over to adaptive voltage mode.  In adaptive mode, your cpu will only go to that maximum voltage and frequency when it needs it, so most of the time it will be at low temperatures and voltage, which will certainly extend your cpu's lifespan.  

If you have any questions on adaptive mode, mention me in a post and I can help you out.  Adaptive setup is really easy, but not very intuitive.  

It adaptive mode required? My idle temperatures are around 35 degress, so I think adaptive mode will not create a big difference. Also, I heard  that adaptive mode can cause voltage spikes that can decrease CPU lifetime. So I am between two options:

 

- Run the CPU in adaptive mode and have a lower voltage, but maybe cause voltage spikes

- Run the CPU in offset/override mode at a constant voltage, low chance of voltage spikes.

 

What do you recommend? 

[msvelev]

My 4690K runs at the same speed/volts. It's better on a 4790K tho

[Queek]

It adaptive mode required? My idle temperatures are around 35 degress, so I think adaptive mode will not create a big difference. Also, I heard  that adaptive mode can cause voltage spikes that can decrease CPU lifetime. So I am between two options:

 

- Run the CPU in adaptive mode and have a lower voltage, but maybe cause voltage spikes

- Run the CPU in offset/override mode at a constant voltage, low chance of voltage spikes.

 

What do you recommend? 

The voltage only spikes in adaptive mode if you run stress testing programs.  Provided that you only use your computer "normally" (e.g. gaming, web browsing, not synthetic tests), your voltage will not spike to abnormal voltages.  I personally have had an adaptive overclock on my 4690k for the last year, and have never seen large voltage spikes because I don't run synthetic tests on adaptive mode.  

Your temperatures are low, sure, but you asked about cpu lifespan.  Turning it on adaptive mode will probably increase your lifespan.  I can personally vouch for adaptive mode being awesome.  

Don't run offset mode, that mode gives some really funky voltages.  Offset mode is very unpredictable (ie. high chance of voltage spikes), by all accounts that I've heard.  

[wiebevb]

The voltage only spikes in adaptive mode if you run stress testing programs.  Provided that you only use your computer "normally" (e.g. gaming, web browsing, not synthetic tests), your voltage will not spike to abnormal voltages.  I personally have had an adaptive overclock on my 4690k for the last year, and have never seen large voltage spikes because I don't run synthetic tests on adaptive mode.  

Your temperatures are low, sure, but you asked about cpu lifespan.  Turning it on adaptive mode will probably increase your lifespan.  I can personally vouch for adaptive mode being awesome.  

Don't run offset mode, that mode gives some really funky voltages.  Offset mode is very unpredictable (ie. high chance of voltage spikes), by all accounts that I've heard.  

Ok, thanks. One more question, should i choose adaptive + offset mode or adaptive mode in the bios?

[Queek]

Ok, thanks. One more question, should i choose adaptive + offset mode or adaptive mode in the bios?

I'm unfamiliar with your bios, but you usually have to set two values:

1) additional turbo mode voltage (that's what the ASUS bios calls it), which is the manual voltage you found from stress testing

2) an offset (the offset cannot be set to 0 in my experience, so set it to +0.001V for the time being)

 

You may run into idle instability in adaptive mode (crashes when you're not doing much), in which case you will have to play around with the offset.  I explain that in detail in this thread.  It may take a couple tries, but once you get it down, you'll be rock-solid stable.  

[incarnate]

It gives me a warm fuzzy feeling knowing that people still remember that article

Unfortunately LLC has become irrelevant, as Haswell and Devil's Canyon voltage regulation happens on the chip itself. LLC now affect Vin only (the voltage supplied to the chip), and I have found that different LLC settings no longer affect stability.

It appear that you have a very nice chip! Congratulations!

You should see how high you can get it, for science. That way those of us with only mediocre chips can live vicariously through you.

Ah, good to know. It was a great guide to llc.

[Br3tt96]

I got 4.6GHz @ 1.20v so far stable 66c max during intel xtu stress test. I had my pc at 4.8GHz @ 1.22v for a a couple of days before it crashed. I went to 1.225v and the temps were some what hot during Intel xtu stress test 71c max temp... So now I'm back at 4.6GHz @1.20v

[wiebevb]

I got 4.6GHz @ 1.20v so far stable 66c max during intel xtu stress test. I had my pc at 4.8GHz @ 1.22v for a a couple of days before it crashed. I went to 1.225v and the temps were some what hot during Intel xtu stress test 71c max temp... So now I'm back at 4.6GHz @1.20v

Maybe I could reach such voltages also, i have not tested that. But I am running stable right now, so no necessary need to change anything. 

[incarnate]

I got 4.6GHz @ 1.20v so far stable 66c max during intel xtu stress test. I had my pc at 4.8GHz @ 1.22v for a a couple of days before it crashed. I went to 1.225v and the temps were some what hot during Intel xtu stress test 71c max temp... So now I'm back at 4.6GHz @1.20v

 

during stress tests, which create an unrealistic heat load, 70 degrees is fine. once you get to 80-85 then it is considered to be getting too hot. i wouldn't worry about 70 degrees during a stress test at all.

[wiebevb]

A little update

 

I tried to but the voltage back step by step and I can boot to 1.17 V at 4.6 GHz and run a stress test for 3 minutes in AIDA64. I know you have to stress much longer for a reliable stable CPU. Currently, i am running at 1.185 V (little bit higher on load, around 1.19 V) and I try that for a couple of days with some additional stress tests.

Two questions: Is it even possible for the 4790K to run at that voltages , and what kind of advantages I can expect?

My temps are around 70 degress celcius on highest core, ambient temp around 24 celcius.

Cooler: NH-D14