Exactly how much does a PSU affect ability to overclock?

Just wondering if anyone has any hard data on exactly how the PSU affects maximum overclock potential. How do you know that your PSU is your bottleneck? Also, is a vcore fluctuaion of +/- .02 volts a serious issue in OC'ing?
11 answers Last reply
More about exactly affect ability overclock
  1. I don't have any solid answers, but I'm DEFINITELY interested in the questions you brought up. I do believe my VCore fluctuates (perhaps more than yours) on my X2-4400 OC'd to 2.6 GHz. I've hit an absolute wall with my OC... I can't got one bit higher... yet at this speed it is 100% stable. I'm inclined to think it's my power supply since it's not a thermal/reliability issue.

    I'll be watching this post! Hope you get some good responses.
  2. I also have hit a wall in my OC. I know it's not a thermal issue or a memory issue. I would really like to know if it would be better to replace my run-of-the-mill PSU with a higher quality one. My system runs stable at 2.8GHz and I would like to up that a bit, but I don't want to spend $120 bucks and then find out that it had no bearing on my system performance.
  3. The main reason it can affect your ability to OC is what they call "Clean power" in simple terms its hmmm this ------- would be a clean line while a cheaper PS's line might look like ---^---^-----^^^---- with spikes and dips :) mother bourds have circutry on them to help filter that out a little but they can only do so much it takes a good mobo combined with a good PS to get clean power :)
  4. Power supplies are not a bottleneck. Having good power clean power is nice for all systems, but the main cpu power comes from voltage regulators on the motherboard, not the PSU. Course, the main problem in oc'ing is not actually the cpu power, but the rest of the system power... unless the buss speeds aren't changes (fsb or htt, whatever) from officially supported speeds. I suppose that is sort of contradictory, and I'm sure someone will disagree with me, but I'm more inclined to believe PSU doesn't have as much to do with overclocking as people want to believe.

    Any fluctuations in vcore voltage you might see in monitoring software should just be ignored. You can not know how accurate or reliable that stuff is unless you break out some test equipment and look at it yourself, and few people have the equipment to do it. All in all, use that monitoring stuff as a general guide and then close it down and never look at it again, because it tells you little.
  5. Vcore fluctations are never good. They tend to happen under load also. As for the affect on your overclock it will have an effect, but I honestly don't know enough to say how much. From what I understand is that as you overclock and raise the vcore your proc is pulling more and more energy from the PSU (fairly easy to see). I think the problem arises from the fluctations while OC'd. The cleaner the power (less fluctuations) the higher the OC, but to what extent idk. We need a PSU expert in here... lol

    What chip are you using sdrawkcabgod? I can at least give you an idea if its a processor limitation, board, or just that your done OC'ing.
  6. Quote:
    ...but I don't want to spend $120 bucks and then find out that it had no bearing on my system performance.

    I'll drink to that!

    Anyways, I bought a "brand name" power supply (Antec) so it's not like I'm expecting miracles out of a PSU that was included with a $40 case.
  7. This might cause more questions than answers but, with my system 3500+ msi k8neo4 platinum corsair xms ram (see sig for more). I can only oc 100mhz before hitting a wall and really not even that because that extra 100mhz my pc can barely boot to windows, run anything and it will crash after a few moments.

    I mess with ram timings, voltage, cpu voltage, memory dividers, everything, no luck. I have great cooling and plenty of power (680W), there are 2 things that I can pinn my problem on, 1. The cpu is just an absolute pos for overclocking (newcastle core I have heard are bad oc ers, but nothing like what I got bad) 2. The vcore fluctuation that ranges from +- 0.05 to +- 0.12

    Every time when my cpu is oc'ed, when the vcore drops the pc craps. I thought that the vcore fluctuation was a bad psu, but of late have learned that is due mostly to mobo not psu as mentioned above.

    So I think that fluctuation has a lot to due with your oc ability, affected by both psu and mobo. If I am wrong about that would someone please give me something to try so maybe I can get more than a 2.3 ghz oc, I mean that is just pathetic, 100mhz.
  8. Yeah your vcore fluctuation makes me nervous. .12v is a pretty big diffrence and probably responsible for your sh!tty overclock. I dont have any personal experiance with the newcastle core.

    If you have put your RAM on a divider <1 and still can't get an OC then the problem probably lies with your board or your proc. 100Mhz is such a small OC that I can't imagine a Newcastle giving out that quickly but then again there are always a few freak chips lol.

    I haven't heard anything terrible about your board. MSI seems to be a solid midrage board. One thing you have to do when overclocking is disable Cool N Qiuet, if you don't your hosed. Also trying leaving most of yoru settings on Auto except voltage and FSB. Drop the HTT multiplier down 1 to 4x. If you still can't get more than 100mhz, I would say your PSU/Mobo combo really can't keep the power clean enough for you.

    Take this witha grain of salt, im speculating a bit.
  9. I cant remeber if the cool n quiet was off or not, Ill try tonight, also I never tried to lower the htt multiplyer then up the fsb and im gonna try that too. thanks for that info, but im not too worried about oc'ing my pc is just fine for me how it is. just agrivated that I am unable to squeze more out of it so I can hold off buying anoter pc for a little longer when it is no longer good enugh for me.

    <edit>btw most of the time vcore only changes by about .05 just randomly every now and again it does a major jump of like .07 and the most I saw .15, but it has crashed when it only diped .01 before so I dont know wtf is going on bc it changes that much at stock 2.2ghz and is rock solid. unless its that cool n quiet thing (crossing fingers hope its something that simple) I think im just sol.</edit>
  10. Give it 6-8 weeks and I think AM2 should be out which will represent a solid time for upgrade if need be. Unless you think you can hold out till Q1 2007 and AMD's 65nm launch or Q3 2006 for Conroe's Launch. Let us know
  11. Oh I have no need to upgrade now I can eaisly wait for amd or intel to establish who's best, then buy. Infact I hope to have my current rig for 2 more years. I play games a good bit so idealy I at most want to upgrade to more ram and better vid card in a year, but thats only if I just have to have eye candy and I got some money to blow.

    I acctualy just want to build a whole new rig every 3 to 3.5 years then keep the old ones as file servers and cram a cuple TB each of hdd space in them so I can make images of all my dvd's (now nearing 150 in my collection so 4.7gb x 150=705gb just right now not to mention how many more Ill have in 2~2.5 years)
Ask a new question

Read More

Power Supplies Overclocking Bottleneck Components