• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

N68C-GS FX -- Power limits?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

ronbaby

New Member
Joined
Apr 6, 2013
I've been building systems from parts for over 25 years, but I've always been too conservative to even try overclocking. Now, at last, I think that I may perhaps give it a try. But I need some advice/guidance.

I've just purchased an Athlon II x3 3.0Ghz CPU and the subject motherboard, i.e. and ASRock N68C-GS FX. (Yes, I _am_ a cheapskate.)

I'd like to try overclocking the Athlon, and maybe also try unlocking the fourth core, but I'm worried about power limits relating to the motherboard. This CPU is already rated at 95W (_before_ overclocking and/or unlocking the fourth core) and the motherboard is only rated for 95W. So what can I get away with here, you know, _without_ burning up this motherboard?

P.S. I'll be using a Thermaltake TR2-R1 heatsink/fan combo (which I think is way overkill for this CPU) if that makes any difference.
 
:welcome: to the forum ronbaby.
That is a very minimal Mobo indeed. I couldn't check to see if the board supports unlocking the core but my advice would be to download Prime95 and HWmonitor(non pro version) and run them both at stock. Keep an eye on your temps don't let the core go over 60 and the socket(cpu) which I think is displayed as tmpin2 on that board below 70. This will give you an idea of weather or not you CAN OC.
 
Thanks much for your quick reply.

This board most definitely _does_ have all of the ASRock BIOS goodies related to overclocking _and_ also it has this thing they call "UCC" which I gather gives me the capability to unlock that fourth core (UCC == Unlock Cpu Cores).

So the bottom line is that the board has the basic capabilities to allow me to do these things, but I am still not sure whether or not, by doing them, I will burn up the board due to excessive power draw by the CPU, beyond what the board was designed for (95W max). So I still need an answer to that question.

Regarding the temps, although CPU temp won't tell me whether on not the _board_ will survive, you know, if it has to supply more than 95W to the (OC'd) CPU, you are certainly correct that I should also and additionally be checking CPU temps as you suggested. But I stress again that the CPU temp, even under load, isn't going to tell me whether or not I'm about to burn up the power traces ON THE BOARD, and this is what I am primarily worried about.

P.S. May I safely assume that all of the temps you cited are expressed in degrees C? (Sorry, but I have to ask. I'm an American, and we still do temps in degrees F, generally.)
 
In forums, most of we Americans do the Centigrade thing since the white papers are not in F but C.

Let us put it this way. To overclock is to run parts out of specification. Now you want a safe out of specification, specification? Wow.

The board should handle the 3 core cpu to 3.7Ghz or so if the cpu and board can play together in doing the adjusting. Note the qualifier in that I said handle but made no mention of that as a speed you will attain.

(Yes, I _am_ a cheapskate.) << is not a methodology that bodes well when then wishing to run your parts out of specification as in overclocking. It is not that you cannot or even perhaps should not but that your results may be less than satisfactory and accomplished with much trial and error at your own keyboard since we that really are into overclocking, buy or purchase with that ambition in mind.
RGone...
 
To overclock is to run parts out of specification. Now you want a safe out of specification, specification? Wow.

Yes, for some value of "safe".

Obviously there are different kinds of things that can and do go wrong when one is overclocking. It is my understanding... and please do correct me if I am wrong... that the most typical result of excessive overclocking is that the CPU (or whatever) in question will simply start to produce incorrect results, and that after this occurs, the clocking of the device can be scaled back down until it is within spec for the given part, and that subsequent to this, there will be no lasting ill effects of the prior attempt to overclock a bit too much.

Obviously, this is quite a bit different from physically burning up either the part in question or the board it sits in.

If you have overclocked a CPU too much, then it is obviously not "safe" to then rely upon what that CPU then tells you is the square root of 724569. Specifically, the result in question should probably not be used in a potentially life-threatening context (e.g. in a pacemaker) or as part of completing your tax return. But this is a different kind of (lack of) "safety" than what I was asking about.

I was asking specifically and only whether or not overclocking my Athlon II X3 and/or unlocking its fourth core, whilst the CPU in question is installed within an N68C-GS FX might cause IRREPARABLE and permanent damage to the hardware itself. And yes, I do believe that, depending on the design of the board in question, it may be possible to (at least) guarantee _this_ kind of "safety" even if the "safety" of relying on the results produced by the CPU under such conditions might not be "safe".
 
1. In general 1.5V to the cpu is "considered" safe. Some say higher and there we go with the various opinions of just what is 'safe'.

2. The greater the Vcore in general the more heat to deal with and the greater the strain on the circuit that supplies power to the CPU and that circiuit is mostly known as the VRM circuit. N68C-GS FX does not appear to have the most robost of VRMs but you should be fine with a 3 core cpu. At least to the 1.5Vcore mentioned in #1 above.

3. Unlocking a core fully DISABLES the ability to read the most important temp within the cpu. So software such as HWMonitor (free version) most of us use or any other software temp monitor will not output a CPU Core temp.

4. From #3 above you should get a baseline of the comparison of the cpu socket temp to the cpu core temp so that you can then add or subtract the difference of the socket temp to cpu core temp and then know the Cpu Core Temp by what the CPU socket temp is reading. It has always been this way when unlocking AMD locked core cpu.

5. It is often better to determine to what rate you can overclock before trying to unlock the core since some cores are locked and need be since they are faulty. If it overclocks fine in locked mode and is a double handfull to overclock with cores unlocked then likely they locked because the core is for real faulty.

6. We normally use Prime 95 used in Blend mode to apply a load to the cpu to give some determination of stability and need for more cpu voltage or if the heat is getting too high when fully loaded as Prime 95 Blend mode tends to do. 70c on the Temp that HWMonitor is showing in the list of TMPINx. No idea which temp that board will readout since it is made using an Nvidia chipset and they have not come to market with new chipset in over 4 or 5 years now. It is a dated chipset and not in great use here in the forums from the dated reasoning. 60c is suggested as a max for CPU Core Temp but in latest versions of HWMonitor (free version) the CPU Core Temp is called "package" temp and will be the one that disappears when you unlock an AMD locked core cpu.

About it I think. Best I can remember.
RGone...
 
OK, without getting too awfully deep into the intricacies of the board's VRM (which are probably too deep for me anyway) allow me to clarify, in simple layman's terms, why I was asking these questions in the first place.

My motherboard is rated for a 95W CPU. End of story. I assume that if I have a CPU in the socket that is drawing more than that, then I run a certain, perhaps modest or perhaps large risk of burning up the board.

Meanwhile, I have an Athlon II X3 3.0GHz which according to AMD is nominally a "95W" CPU. But I harbor more than a little doubt about that number, and I suspect that in reality, for this specific CPU, the actual nominal wattage is lower, even at rated (non-overclocked) speed and maximum load.

Why do I say this? Two reasons.

First, I cannot help but note that essentially every Athlon II CPU that ADM produces has been rated by them as being either 65W or 95W or 125W. Given the many variations of this basic CPU that they produce, the fact that there are only three wattages specified for every member of the family seems on the face of it to be clearly NOT reflective of the actual power draw of all of the specific different CPUs in this family, but rather, and more likely, these three wattage ratings are just a dumbed down simplification, suitable for marketing purposes and for the typical end luser community (who do not generally benefit from being confused by the actual facts).

Second and more specifically, I have noted with more than passing interest that the specific model of this family which is an Athlon II *X4* 3.0Ghz CPU is also rated by AMD as a... wait for it... "95W" CPU. I mean huh? WTF?

In short if we blindly believed the wattage numbers that AMD is promulgating for these CPUs, then we would have to also accept as truth the notion that one could go from an X3 processor to an X4 processor, within the same family, and at exactly the same clock speed, and thereby get something up to another 33% of work done absolutely FOR FREE in terms of additional power draw.

This is obviously nonsensical. There are plenty of ways of obtaining additional compute horsepower for a small amount of additional power draw, but as far as I know, there are no practical scenarios under which one can get additional compute horsepower at a cost of exactly -zero- additional watts of input. Not without substantially re-engineering the entire processor, at least.

So I think that you can see where I am going with this. What I would really like to know is: What is the *real* maximum power draw of an Athlon II X3 3.0Ghz CPU? Common sense says that despite what AMD says, it is something less than 95 watts. If it is in reality, say, 86 watts, then I personally will feel comfortable overclocking this CPU by 10% and would do so without any serious fear of causing my motherboard to break a sweat, i.e. by making it have to handle more than 95 watts of flow to the CPU socket.

Perhaps I am being naive in this rather simplistic view of things, and it seems to me probable that there are many many potential complications and additional factors that I have not accounted for, so much so so that someone might be able to write a PhD dissertation on the whole question. But I am a simple man and I like simple answers and I don't really have either the time or the inclination or the motivation to delve too awfully deeply into whether or not I may in fact end up burning up a $45(USD) board based on my simplified assumptions. So for now, I just simply want to know if anyone here happens to know the *actual* max power draw of an Athlon II X3 3.0GHz. That information would be helpful. And if I do burn up the board due to my over-simplistic assumptions, then I promise to hang my head in shame and report back here, nonetheless, that I have done so, for the benefit of the community.
 
Well you are digging too deep. In fact you are digging deeper than AMD or Intel either one is willing to talk much about. You can for your own pursuits, find out what both AMD and Intel say about their TDP ratings. Then search a relatively new term AMD is bouncing around called ACP. You want double-speak? You got it with that one for sure.

However in coining the ACP mindset, AMD says they finally took the time to break into a cpu and measure some actual power draws. Well just whoopee. They then break out a chart that shows a TDP rating that corresponds to a newer ACP rating. This was done sometime ago and is more pointed to your type cpu but certainly not the new tech FX processors. But you also need to realize that in truth neither AMD nor Intel are overly interested in the consumer Cpu's. They are more interersted in the big money area of servers for business. So most of what you will actually be able to fihd is only going to relate to business type processors. I understand the why of that. Too many who don't know jack causing rumors.

After you get thoroughly ticked trying to cross reference info that you can find and making it relate to the consumer cpus, you can do what most of us do in the course of overclocking. We have seen what most get by with doing. We inch our way toward the results others have gotten and are glad we also can do simillar.

When I sat at the keyboard and wrote out those 6 points in post #6 above it was not without having done a good deal of overclocking of AMD and Intel cpus. It was after having managed a motherboard company forum backed by the company and when those of us whom tried to help did so, it was with an eye to being sensible.

When I wrote this earlier:
2. The greater the Vcore in general the more heat to deal with and the greater the strain on the circuit that supplies power to the CPU and that circiuit is mostly known as the VRM circuit. N68C-GS FX does not appear to have the most robost of VRMs but you should be fine with a 3 core cpu. At least to the 1.5Vcore mentioned in #1 above.
It was not intended to be nor was it what probably 1,000s of us would consider to be anything but conservative if we have been into overclocking as long as some of us have been.

Now I have not minded a back and forth communication process at all with you sir. But I tell you honestly that you have gotten to the point of asking questions beyond what only people like AMD or Intel engineers could answer had they done real testing of the actual power draw of each of their processors and not couched their power draw in terms lke TDP. Then know that they will not tell anything beyond what each company will release in their white papers or perhaps lose their jobs.

And while you are looking up the various TDP information keep your eye open for that phrase that is in the AMD and the Intel papers where it will be said that TDP does NOT reflect the maximum current their processors can draw. That one statement is in both Intel and AMD literature. I thought it was a plain 'hoot' to read the phrase in both companies white papers about TDP. We run all over the world trying to get TDP settled in our minds and then that phrase. What a riot!

To overclock anything from a computer cpu to the cpus in cell phones is to walk up and shoulder the risk, small or large, of running our parts outside the original specifications. If one cannot shoulder that burden then they do not need to even worry with overclocking.

I had a borrowed motherboard and cpu loaned to me by a forum member. He was very busy and without the time to put the cpu and that motherboard thru their paces and write up what I found in a thread or two in the forum.

I put his FX-6300 processor in my own motherboard first and just was not willing to push my board hard enough to run his cpu at 5.5Ghz. So he said put HIS dang cpu in his motherboard and get a CPUz Validation at 5.5Ghz. I said 'yessir ree bob' and put his cpu into his motherboard and put the pedal to the metal. He got his 5.5Ghz cpu validation of his FX-6300 at 1.71Volts to the cpu on his Asrock Fatal1ty 990FX motherboard. I think that thread was titled "is it a contender" in this mortherboard forum secition for AMD motherboards.

Point being he bore up to the risk and said to me to do it. I know him and he would not be ticked at me for blowing his shett up if it had happened. But in this overclocking thing, you have to just walk up to the door and open it and walk thru. No manufacturer is going to speak much about running their product out of spec and will not say much about the regular consumer cpu anyway. They are really interested in the big money server stuff.

Enjoy yourself and leave off quite so much questioning. Learn some on the cheap board and cheap processor. Take sensible baby steps and monitor those aspects I spoke of in post #6 since it was aimed dead at the newer person trying to begin to overclock on a rather cheap motherboard. Good luck sir.
RGone...
 
Thank you for your detailed reply.

It is clear to me after reading what you wrote that you're a person with some serious experience in this area (overclocking generally) and although I may do a lot of dumb things, I like to flatter myself that I'm at least smart enough to take good advice from folks who are obviously my betters.

So I shall. I'll take your advice, read some more about how to do this stuff... I don't even really know the basic procedures yet... and I'll do what you said, i.e. take baby steps, see what works, and be prepared for possible failure up to and including possibly burning up my components. I can live with that. Neither the CPU nor the board really cost that much anyway, after all. (These are *not* by any stretch of the imagination "high end" components.)

It will certainly be educational. In fact it already has been.

Thanks again.
 
Honestly "gabrimor" you can put the VGA sinks on those 6 mosfets, but I am not sure the money is worth it on that very cheap motherboard. The VRM for the ram is over just above the 24pin ATX power connector and the Ram VRM is usually included in the power circuit description and is why you see 4 + 1 or 8 +2 in that the 1 or 2 indicates the memory VRM circuit. Putting heatsinks on the 6 mosfets above the cpu socket might help some. Air passnng thru or over the VGA sinks is probably just as critical.

Myself I would probably use the board as it is designed to be used and forget overclocking and save the VGA sink money to go on a much more robust motherboard in the future. Beyond that is whatever you want to do since the money is yours to spend as you wish. Luck man.
RGone...

EDIT:
I went to OCN where they are keeping a fairly up to date listing of the VRM circuits for the AMD motherboards. You can follow the link below and find that N68 mobo and the suggestion they give about an overclock of that mobo and the fact it has only 3 phases for the cpu power circuit.
Link >> https://docs.google.com/spreadsheet...UFlMEVWeFhuckJEVF9aMmtpUFE&chrome=false&gid=2
END EDIT.
 
Last edited:
Back