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

SOLVED Finally had to give it a go...FX-8120 to 4.8Ghz

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
frakk, from what i have seen of actual screenshots i think that you are right.
but you have to dig through a lot of forums to find posted screen shots.
Rgone makes me wonder where he got his.......

Your just hell bent on him cheating somehow hehehe... :p

They do seem good temps for that clock. but look at mine compared to most Thuban's.
frakk, he lives in a place that is so humid and hot it's nick name is mississauna.
he must have his computer in the river not next to it.
strange thing with your numbers is that they are so good, but one guy on extreme systems has numbers close to yours but with a vcore below 1.4, i am amazed how cpu's differ one to the next........
It is, we need a way of predicting it, if that we're possible

c/o recycle bin ?
Hey RGone... I know you where kinda sweating out the voltage increases and I can say that when I changed to a motherboard that had the Digi+ VRM I was able to cut .05v or more from my OC after playing with the settings. So I would recommend going into the Digi+ controls and playing with that you may be able to get that voltage back down to a lower level.

Also it already sounds like you got the idea of cooling down the vrm and socket which is super smart as I had posted a chart a long time ago showing that for every 2 or 3 degrees the VRM dropped the CPU dropped 1 and that was even on a WC setup so that helps big time. I found it best to have 1 fan behind the MOBO and one over the top of the VRM.

Those are nice numbers I almost want to go to a BD from my 960T but I was blessed with a 4GHz x6 stable 960 at a decent voltage. Like I said though def try out those Digi+ VRM controls.
@"givemedew", I have a fan blowing across the hole I cut in the back of my bench motherboard tray, so it draws the air away from the back of the mobo. Still have the fan blowing on the VRMs.

Have played a 'little' with the Digi settings. The thing I found that surely drops Vcore required and thus temps is the HT and CPU/NB frequency. Good for about 1.5C and about 0.015 Vcore.

Yes, I really like the CHV because it has adjustments for days.

After I get Frakksters 4.8Ghz, I am going to slow it some and give 'all' the settings a testing to see what moves what around. Thanks for the heads-up. RGone...ster.
For the Frakk...ster.

4.75 was not enough for the Frakk...ster, so here is the 4.8Ghz he made mention of. Since this time I did n0t make a silly 200mhz cpu speed jump it was much easier to dial in the Vcore. Two bumps in the CHV VCore bios setting was what it took. In bios it is set to 1.5825 and the CPU LLC is set to "Ultra High" and when loaded the Vcore jumped to an indicated 1.57 volts to run the 4.8Ghz.

This is a used FX-8120 that has been used pretty hard before I got it. I would not say abused at all but surely used in overclocked mode. Probably at the speeds I have reached, my voltage is less than the cpu has seen before and likely due to running a little cooler.

I have looked at AMD's own overclocking guide and in that guide they give some voltages that 'could' in the correct circumstances be considered safe even if overclocking voids the cpu warranty. Hehehe. Heard that. So with the voltage numbers that AMD gives under very extreme circumstances with Gi-normous cooling, I don't feel too bad using a little over 1.5 Vcore on the cpu.

However it must be understood, that this is not my very first rodeo. Been at this stuff for years now. I know it is all a chance and the big boom could come at any minute. But I make calculated guesstimations based on experience and hours and hours of searching for what others do and how. So my voltage with my cooling seems okay to me for today. YMMV greatly.

I am being a little windy here since putting as much circumstantial information in as possible, will allow this thread to be linked to others perhaps and save some of this constant writing. Most come into forums wanting a gimme overclock and are gone when they get what they came for if possible. A lot of us hang and spend hours typing. I hope to save myself some typing by making this thread.

Anyway, 1.57 Vcore bumped in by CPU LLC when loaded, did the deed to reach 4.8Ghz and pass OCCT P95 testing for 2 hours.

On note, this rig normally runs with Cool N Quiet enabled as well as C1E and for myself, I have no issues related to stability. This CHV responds immediately with voltage to jump from 1.5Ghz to ~4.5Ghz in a heartbeat. I remembered this time to disable 'balanced power mode' in Winders before running the 4.8 test so I could have the cpu stay throttled up for the screen capture at the conclusion of 2 hours run of OCCT.

For you Frakk...ster man.


  • 8120 at 4.8Ghz OCCT.jpg
    8120 at 4.8Ghz OCCT.jpg
    156.6 KB · Views: 100
  • 2012-09-02-22h34-Memory Used.png
    2012-09-02-22h34-Memory Used.png
    18.5 KB · Views: 100
Alright, i'm happy now :D Thank you :thup:

Do you have a link for AMD's overclocking guide? might be an interesting read.
That is a nice OC ..... and some pretty temps you have there. All joking side that must be some outstanding cooler you have. I am soon headng home and will give this tread a better look see on my big screen. Dam I don't mind this little ol laptop but the screen and my eyes don't really mix anymore ...... old age I guess.
Alright, i'm happy now :D Thank you :thup:

Do you have a link for AMD's overclocking guide? might be an interesting read.

This guide is in .pdf format. They make some pretty good definitions of some of the names given to bios menu items and are pretty clear on warranty disclaimers. Nothing says I 'must' follow their procedures since I don't overclock with software, but the how and why they are testing the clocks and what is adjusted to get stable, are universal ideas. I don't have to DO IT like AMD does it but there is nothing wrong with some of the background information they share.

I am afraid many read something wanting to hear and see what they 'wish' to be true. I don't find that to be very useful, since rumors seldom prove to be true and what I want is seldom how it is. Especially so with overclocking. It takes time, money and effort to really overclock for long term use.

I did find the AMD guide to be useful for day to day information that I did incorporate into my knowledge of what to expect from the Bulldozer.

AMD FX Processors Unleashed | a Guide to
Performance Tuning with AMD OverDrive
and the new AMD FX Processors

By the way AMD has such guides for their processors dating back at least two processor/chipset series if I am not mistaken. I think I even glanced at such a guide for the FM1 APUs.
RGonester, can't find it now but somewhere yesterday in one of your posts you made mention of "some memroy latches being stouter than others" especially 1:4 being better than 3:10. I'm curious about that. What did you mean by "stouter"? Better performance or allowing higher overclocking? Can you elaborate on that?
"trents", I did say something along those lines and said I had glimpsed something along those lines but was not sure myself but set mine so that it was 1:4. So I used that thought of 1:4 in my setup, for a 'best' performance idea, but was not sure if such was really so. I don't imagine it could make a huge difference.

I have got to take my Dad towards New Orleans since they have electricity again at my sister's home where he lives. I will do some more digging when I get home later today.

Oh I like even dividing since all the speeds derive from 200 it is more accurate to use 1:4 dividing from 200. Hehehe. That though is just me. Will try to find where I glimpsed that 1:4 thing. It might be just so much BS or neither good nor bad. Later RGone...ster.
RGone I don't think you are the only one that has said that about the memmory divider as I remember reading it here somewhere too. It wasn't in a thread of its ownbut burried in another thread here somewhere but for the life of me I can't remember where either.
RGone I don't think you are the only one that has said that about the memmory divider as I remember reading it here somewhere too. It wasn't in a thread of its ownbut burried in another thread here somewhere but for the life of me I can't remember where either.

Yeah, I did some checking and I think the idea may be just BS. I am doing some more checking, but I don't see how you could choose any particular memory divider and maintain it thru all ranges without the ram and CPU/NB or another bus becoming overclocked at some point and then any divider might seem better than the other.

So that I did say it earlier is not untrue; since I did say it. However I could use the 1:4 but did it give greater performance somehow? I am now having doubts about that being accurate. I need to do some serious juggling and testing. RGone...ster.
I trust your trip to the Big Easy was safe and uneventful.

Yes, got my Dad home safely. Thanks for asking man. Actually I did not have to go all the way to N. O. Groceries are still in short supply in N. O.. so my sister came to a Walmarts just inside the southern MS border to buy groceries and we only had to go halfway as did they. RGone...ster.
Shooting for fastest Mhz at 1.34Vcore or less...

I think we often confuse Max CPU speed with goodness and not a more delicate balance for speed and without the heat of the current FX-series of processors.

Although I have mighty good cooling, I can interpolate by voltage what would be a great reduciton in temps involved with the FX type processors when nothing but running it WFO seems to do most. I am reversing that idea slightly in that I want to find what the Max CPU Mhz I can attain with 1.34Vcore or less. That would put people; with CM 212 EVOs or very similar, into a temp zone that is fully controllable without having to resort to 5 or 6 case fans to remove the heat from the case or even to have to go with some sort of W/Cooling to handle that other 200 or 300Mhz just to run WFO.

The Capture below is of ~4.0Ghz with 1.31Vcore we will round it up to 1.31V even though 75% of the time after OCCT was started and running for more than 5 mins...the Vcore dropped to 1.3V. This most likely due to CPU LLC being set now to "High" and not "Ultra". I am wanting to use as little VCore as possible in the testing now.

Cutting the Vcore as low as possible over a number of mini-trial runs, I found that for me I could go ahead and "drop" the multiplier and ADD FSB/HTT "CPU Frequency and the Vcore needed was lessened by about 0.06. If one were rounding that up like most applications tend to do that is nearly a tenth of a volt. Which when looking for the very minimum of Vcore, well every little bit helps.



  • 8120 at 4Ghz 1.31V.jpg
    8120 at 4Ghz 1.31V.jpg
    185.8 KB · Views: 70