4 Replies Latest reply on Feb 17, 2011 12:34 AM by tomr

    "overpowering" of CPU

    tomr

      Hello everybody, may I ask you one "simple" question. I´ve asked it few times before, but I still didn´t get a clear answer :-( So I hope, that here I will get my answer since I believe, I am at the best place to ask :-)

       

      I purchased SB 2500K, which is meant for overclocking and get these results:

       

      Intel burn test was stable (20 runs) on 4.5 GHz ~ 1.305 V
      However Prime95 crashed in few second. 4.5 GHz in Prime95 was stable (22h 30min of running) on 1.345 V (max temp. when testing was 56°C).

       

      Some friends of mine told me, that "overpowering of CPU" may cause harm to it or even reduce its lifetime. On the other hand, some insured me, that I shouldn´t worry about it... However, I am new for overclocking and I would finally appreciate some experienced advice very much... :-(

      Most of time (like 95%) my CPU frequency drop automatically to 1.6 GHz (because of ASUS motherboard power management), BUT voltage reamains on 1.345 ALL THE TIME!

      So result is, that CPU temperature is below 30°C, but it is in 95% of time provided by far much more voltage, then it actually need.

       

      So my question is really simple...

       

      -> can high voltage itself harm or reduce lifetime of CPU (even if temperature sustain below 30°C permanently)??

      You see, temperature is not problem at all, since CPU is not stressed (and even if it is, I can keep it under 56°C thanks to my cooling system).

       

      PS: in the extreme way of thinking (only for my better understanding) - what if I would rise voltage of 2500K for example to 1.5 V on 4.5 GHz, but CPU would still stay below 30°C? Is voltage itself problem to solve or I shouldn´t worry about it?

      (You see... there will be logically some unused voltage anyway most of time on each PC, since nobody stress CPU 24/7... )

       

      I am sorry for maybe banal question, but as a new overclocker I don´t understand principles of CPU power management a lot yet. So I hope I get some lesson here :-) Thank you very much for possible explanation or this problematic and any experienced answer!!!

        • 1. Re: "overpowering" of CPU
          parsec

          tomr, Few questions are banal, if they have good intent, as your's does.

           

          Question: What are you using to monitor the CPU's voltage?  Depending upon what that is, you may not be getting a real reading, you might just be seeing the maximum value if you set that as such in the BIOS.

           

          An FYI, the reason your CPU runs at 1.6GHz is more likely due to Intel's power saving features, those being Enhanced Intel Speed Step (EIST) and C-State technology.  Both of those should be options in your BIOS, to enable/disable, and are likely defaulted to Enabled.  When the CPU load is low, usually the CPU multiplier is reduced, and the frequency drops accordingly.  As soon as demand is put on the CPU, the multiplier will switch back to it's usual value.  C-State's results are similar to EIST, but their actions are more complex and have varying levels or "states" the CPU will be set to in order to save energy.

           

          As you know, CPU temperature is the main thing that damages a CPU.  Frankly, I can't say what exactly would happen if somehow a CPU could be kept at 30C with five volts applied to it, but I would think the extremely tiny transistors would be destroyed by the amount of current that would try to flow through them.  I know that CPUs will throttle themselves down if they become to hot, but I don't know if they can control excessive voltages.  Your voltage, while high, is not beyond what over clocking a CPU would require.  Your question has me wondering too, is it simply high temperature that ruins a CPU, or can high voltage do that as well, independent of temperature.  What we may all be calling "heat damage" might also include over-voltage damage.

           

          While we can discuss a CPU as a single entity, it really is made up of multiple components, including the actual processing cores, cache memory, memory controller, and now, a GPU on the same die.  So while one part might survive a higher voltage, another such as the Integrated Memory Controller, will not.  Regardless, if one part expires, the CPU is destroyed.

           

          I find it hard to believe your voltage is not changing, which is why I need to know what you are using to monitor it.

          • 2. Re: "overpowering" of CPU
            tomr

            Hello Parsec, thank you for your respond. I use CPU-Z (frequency and voltage monitoring) and RealTemp (temperature monitoring). ASUS monitoring tool seems to me inacurate since it shows 10°C lower than RealTemp does (?)

             

            Well, I am glad you are interested in the answer of my question as well :-) So, if anybody knows... May higher voltage ITSELF damage or reduce lifetime of CPU? And if yes, how?

            + if yes => is 1.345 V for 2500K too much?

             

            Thank you for your answers!!

            (I hope somebody from Intel may be helpfull very much in this matter since 2500K is you "baby" :-)

            • 3. Re: "overpowering" of CPU
              parsec

              Ok, CPU-Z does update it's readings, so if you're at 1.345V  constantly, then your BIOS must be set for manual selection of CPU  voltage, and is set to 1.345V.

               

              The problem with wondering if a certain voltage is  dangerous is that these CPUs, and all CPUs for that matter, are much  more complex devices than most people realize when considering power  factors, and have all kinds of things happening that most of us don't  understand, including myself.  If you examine the technical documents  for your CPU, you quickly see things are far more complex than a simple  voltage value.  If we can derive a maximum voltage from the Voltage Identification Definition table, that would be just over 1.5V.

               

              Regardless, the technical documents do focus on the  temperature of the CPU as the main factor to consider for keeping a CPU  from being destroyed.  Related to that, it is not simply the voltage  applied to the CPU (within normal limits) but the actual power in Watts  being used by the CPU, which involves the current draw.

               

              I'm  rambling here, since I am not comfortable answering your question about  your voltage.  Over-clocking and the higher, constant voltages required  for it, are simply the risks that people doing this must deal with.  If  your mother boards BIOS has limits for the CPU voltage that are within a  relatively safe range, then as long as your CPU cooling is adequate,  you are probably fine.  But as with anything operated beyond it's  standard specs, it's longevity will be much more at risk.

               

              Also understand that no one wants to tell you your voltage is fine, and then later your CPU burns out.

              • 4. Re: "overpowering" of CPU
                tomr

                Hello Parsec, thank you for your reply. So I will understand it, that my voltage is safe, since it doesn´t go higher than 1.5V and my temperature is below 56°C. Thank you very much!