Asus m5a97 Pro and fx-8120 OC potential?

[manu2b]

Hi all,

I have an Asus m5a97 Pro and a FX-8120 on their way (I should get them by Thursday).

What kind of OC can I expect, assuming it is an average chip?

Is this board a huge limiter(I'll put a high speed 120mm fan over the VRM heatsink)?

Thank you for your input.

 

[Shadow300]

Not sure about the overclocking potential of this board but it does have a feature where if your clock is too high and it doesn't post, it will be able to reset its settings back to where it can post. The 8120 is a fantastic OC chip as the 8150 ( not far from this one) broke the world record clock speed. Of course that was with extreme cooling, so I'm not sure how far you want to go, but as long as you have good cooling, the board should be fine for OCing since it seams designed for great performance. Just watch the temps on the CPU, they get pretty hot.

 

[mjw21a]

Yup, it does indeed get pretty toasty. You'll most likely be limited by your cooling rather than the board unless you go with proper water cooling. I'd guess depending on your ambient temps anywhere from 4 to 4.5GHz should be fine on high end air.

 

[manu2b]

Thank you both.

Regarding the cooling, I'll use the loop in my sig.

 

[mjw21a]

Ha, serves me right for not reading your sig earlier

 

[Woomack]

@manu2b, Most FX8120/50 are overclocking the same so how high can you oc depends mainly from motherboard and cooling. You can see difference between chips above ~6.5GHz on LN2 but under 5GHz almost all will act the same +/- 0.05V ( that depends from vdroop, cooling etc ).
Most of these cpus are running at about 4.5GHz on good air or more standard water like all these corsair sets and about 4.0-4.1GHz on stock voltage.
In your case it can be something about these 4.5GHz.

 

[mjw21a]

Woah, you've got two very sweet rigs there in #'s 1 & 2 Woomack. How do you find they compare in day to day useage?

 

[manu2b]

@manu2b, Most FX8120/50 are overclocking the same so how high can you oc depends mainly from motherboard and cooling. You can see difference between chips above ~6.5GHz on LN2 but under 5GHz almost all will act the same +/- 0.05V ( that depends from vdroop, cooling etc ).
Most of these cpus are running at about 4.5GHz on good air or more standard water like all these corsair sets and about 4.0-4.1GHz on stock voltage.
In your case it can be something about these 4.5GHz.

Thanks,

With the 360 rad and 3x1800rpm fans (CPU only), I'll aim@5GHz!

 

[Woomack]

@mjw21a, #1 is my daily now. I switched it with my brother who got #2. For games both are almost the same. I'm using single 24" 1920x1080, my bro 25.5" 1920x1200 and most games are running on max details. Some days ago I moved with benching and other stuff to #3 that is based on Z77 board but I'm waiting for Ivy so for now there is Celly 530 ...
Sorry for OT

@manu2b good luck with that but I don't think that you make it stable @5GHz on water ... still have fun trying maybe you make it

 

[mjw21a]

^^^ Yeah, I'm considering a switch to Intel for my main rig but not until Intel have sorted out their heat issues with IB. Having said that, if PD is good enough then it will probably keep me with AMD. If they achieve 10% improvement in IPC and it clocks through the roof then yeah, I'll stick with AMD.

 

[Frakk]

The Mosfet / NB gets warm on those, but as long as you can keep good air flow over those blue heatsinks it will be fine.

I maneged a 4.3Ghz Overclock on my x6, its just my CPU cooler that can't get the heat off the CPU, otherwise it would be fine.

@ mjw21a, one thing about PD that i'm pretty sure off is that they should Overclock pretty bloody high.... even the Stock clocks will be 4Ghz +

 

[mjw21a]

I've a sneaking suspicion that PD may surpass Intel's best, unless they can sort out their heat issues soonish.

 

[Frakk]

I've a sneaking suspicion that PD may surpass Intel's best, unless they can sort out their heat issues soonish.

The IPC on Trinity (based on PD) is 10+ % higher than BD, couple that with the high clock speeds and what you should have there is something that easily beats Phenom II.
It will be interesting to see if that will materialise.

 

[mjw21a]

Yeah, looks like I need to wait for another year before I upgrade again..... See if Intel can fix IB and see if PD improves as much as they look they will

 

[manu2b]

Yep, actually I'll keep both the 8120 and the 955.

If PD is that good, I'll sell the 8120 and get one (should work on a 970 chipset no?).

If not, the BD and the m5a97 will go to ebay/classified, and it will be 955 and IB...

 

[mjw21a]

If PD is sweet I'll use it in my current main, and build a mATX system with a Silverstone TJ08-E for LANs to put my FX-8120 into

 

[manu2b]

Still messing around:
it boots into windows @5GHz/1.46v and cab do main office tasks, but not stable under prime (illegal sumout on cores 6 and 7)

 

[Frakk]

Still messing around:
it boots into windows @5GHz/1.46v and cab do main office tasks, but not stable under prime (illegal sumout on cores 6 and 7)

Can you get a higher vCore? it might just need that little more...

 

[manu2b]

yes, but from 1.52 and up, it freezes 20/30 secs after passing the logon screen (temps at idle around 28°C).

I have to understand what all the bunch of voltage settings mean and how they impact stability.

 

[Frakk]

Other then the usual you have,-

(CPU load line calibration) That calibrates the vCore, when the CPU is stressed it increases the vCore to compensate for vDroop.

(CPU NB load line calibration) as above but with the NB

(CPU current compatibility) I don't know exactly but i think this controls the voltage stability on the mosfet side, again when stressed it gives it extra volts.

(CPU current compatibility) same, this time for the CPU's DRAM controller

(CPU voltage frequency) you can set a pre programmed vCore depending on what speed you have the FSB set at.....

There are endless settings options, it looks like far to many and that its over complicated.
But actually its simply that every conceivable aspect available for overclocking and in that to fine tune to absolute stability at the best thermal solution is there.

Once you start playing with it and then test to see the results of messing about in HWMonitor it all starts to make sense and fall into place.

For example if your overclock is perfectly stable at lower vCore at idle then set those lower volts by using the (Manual Mode) on the CPU & NB Volts option instead of the (offset increase) in the BIOS.
And then use the (CPU load line calibration) setting to boost the vCore once the CPU is loaded up.

My CPU is completely stable at idle @ 1.36v, under stress it needs 1.38v, so i set it (Manual mode) 1.36v and then the (CPU load line calibration) to ultra to compensate for the vDroop and also push it to 1.38v when it needs it, it works remarkably well...
I tested it in HWMonitor, it idles at around 1.36v when stressed it climbs to 1.38v and holds it there..

You can do this for the mosfets, the RAM controller, the FSB, the NB............

You have to play with it a lot to see how it behaves and then use the settings to change that behaviour to get it to behave how you want it to.