Help Me Overclock AMD FX 4130 Properly.

[Lazermonkey]

*Update* I bought a new board. I got the Asrock 970 Extreme4 for $60 brand new.

So far so good!

I recently put together an AMD system.

Specs are as follow.

MSI 970A G43 motherboard

AMD FX-4130 w/ Cooler Master Hyper 212 plus 120mm fan.

Crucial Ballstix Sport 8GB @ 1600mhz

Graphics is an MSI GTX 560 Ti 448 core OC Power Edition. (currently OC'd to 840 mhz from 750 stock) and is stable with very nice temps in the 50's in game and low to mid 30's at idle.

Antec EA-650 watt PSU

Case is an MSI Interceptor Ravager with two 120mm Intake fans. One side, one front.

Three 120mm Exhaust fans. One rear exhaust, two top exhaust.

I currently have it clocked to 4.2 ghz and stable.

But, I am looking to get more out of this beast without damaging the system.

I am not scared to increase voltage etc. just looking for a little guidance in doing so.

TIA for any and all input

 

[Lazermonkey]

No one has any advice or info they can share?

 

[erocker]

It's very simple.

Bump up the CPU multiplier and test for stability at 4.3ghz. If stable, repeat. If not, add a bump of voltage until it is stable. Repeat if you want to go further.

 

[Lazermonkey]

Thank you.

This is my first AMD build and I was used to Intel systems.

Just don't want to fry my new chip lol

 

[erocker]

If you OC'd on an Intel K series chip (unlocked), it's just as easy with AMD.

 

[Lazermonkey]

If you OC'd on an Intel K series chip (unlocked), it's just as easy with AMD.

Thank you.

I am just trying to be safe here.

I appreciate your response

 

[HammerON]

Make sure you use LinX or another program to load your CPU and then carefully watch your temps. Need to stress your system after over clocking to make sure it is stable

 

[d1nky]

I find it simplest to have my clocks stock....set my vcore to the max I want it (check temps at full load)......then adjust the multi with 30min testing between increments, till it becomes unstable. when unstable go back down one and test for longer and then adjust voltage for perfection. test test test

and then theres the LLC etc.....

 

[Lazermonkey]

Okay I have this thing running stable @ 4.3 with stock voltages running Prime95.

One thing that is getting me, though is this one temp in CPUID HW.

The TMPIN1 temp is getting in the 80's range, I am wondering if it is a sensor for the VRM's? or a sensor close to them? the VRM chips on this board are not heatsinked. Not sure why MSI cheaped out when they sinked the two boards above this one? I am guessing it's because they added the USB 3.0 front header on this one and left it out on the boards above this one.

Here is a screenshot of my temps. I hope I am reading them right?

Also. that shot shows my system at 4.2 but it is at 4.3 and steady with the same temps.

 

[d1nky]

try hwinfo64 or coretemp etc, I found with hardware monitor it would literally make up new temp sensors.

its always best to test a few monitors against one another, in order to test the reliability of them.

and by stock voltages do you mean auto? because stock has a wide range, I think its from 1.30 - 1.45?! and your highest 1.440 could get a higher frequency im sure.

 

[Lazermonkey]

try hwinfo64 or coretemp etc, I found with hardware monitor it would literally make up new temp sensors.

its always best to test a few monitors against one another, in order to test the reliability of them.

and by stock voltages do you mean auto? because stock has a wide range, I think its from 1.30 - 1.45?! and your highest 1.440 could get a higher frequency im sure.

Yes, auto.

I have tried Coretemp, Realtemp, HWINFO64 and some others and they are all showing the high temp on the TMPIN1.

I really am perplexed as to what that one is?

One thing I am going to do is buy some heatsinks for the VRM chips as this board soesn't have them.

 

[erocker]

Yes, auto.

You don't want to do that when overclocking. Best thing to do right now is to put the CPU back at stock clocks and read the voltage under load. Then take that voltage (your actual stock voltage) in input that into the bios. Then overclock from there.

 

[newtekie1]

One thing I am going to do is buy some heatsinks for the VRM chips as this board soesn't have them.

I was just going to say, before you start upping the voltage to the CPU put heatsinks on the VRMs. They get hot even at stock, and I wouldn't be surprised if that high sensor is located in the VRM area of the board.

 

[Lazermonkey]

Okay I set the clock back to stock and ran Prime95.

My TMPIN1 maxed at 69c and my voltage peaked at 1.448

My CPU never went past 48c

 

[terrastrife]

You don't want to do that when overclocking. Best thing to do right now is to put the CPU back at stock clocks and read the voltage under load. Then take that voltage (your actual stock voltage) in input that into the bios. Then overclock from there.

If you bothered to look at the screenshot, you will notice that the voltages ARE the standard FX voltages (looks like LLC is enabled and power saving is off however).

Op just play with your multiplier, not your ref bus. Your Packagetemp (coretemp) just make sure that doesn't go over 62C. But yes, without VRM cooling your motherboard won't last very long.

 

[de.das.dude]

i would advise you to get some heatsinks for your motherboard VRM first. that board is kind of famous for burning out its VRM under overclocking

 

[Lazermonkey]

I did not see anything about LLC in my Bios.

i would advise you to get some heatsinks for your motherboard VRM first. that board is kind of famous for burning out its VRM under overclocking

I am going to do that when I figure out which size my chips are so I do not buy the wrong heatsinks

Here is a pic of my Motherboard.

Do these appear to be the right size copper sinks for those chips?

 

[d1nky]

tbh I wouldn't waste the expense, just done some research in your motherboard and well DDD mentioned the vrms burn out, that's because it comes with a massive warning not overclock.

https://spreadsheets.google.com/spr...D79Joo7tdE9xMUFlMEVWeFhuckJEVF9aMmtpUFE&gid=0

get a new board

oh and when overclocking my 4100 I tend to have auto voltages up to 4.4ghz stable (it keeps the temps down on the vrms), and higher frequencies I need to lock on higher voltages and adjust LLC accordingly. but my board is crap and 4+1 phase design and throttles when vrms are hot, it can keep full load and full voltage up to 30 seconds (yes I even timed how long throttling takes for benching purposes)

more voltage and more LLC = higher vrm temps and when full load....itll burn ya finger on the vrms.

 

[erocker]

If you bothered to look at the screenshot, you will notice that the voltages ARE the standard FX voltages (looks like LLC is enabled and power saving is off however).

Op just play with your multiplier, not your ref bus. Your Packagetemp (coretemp) just make sure that doesn't go over 62C. But yes, without VRM cooling your motherboard won't last very long.

I did bother to look at the screenshot. Are you sure that it IS the standard voltage? My FX chip didn't have that high of voltage.

*Lazermonkey, please be sure to edit your posts instead of double/triple posting. Thanks.

 

[GLD]

Some sweet sinks can be found here:
http://www.moddiy.com/categories/Heatsinks-&-Cooling/?sort=bestselling

I have not ordered from here yet, fyi.

 

[newtekie1]

I did bother to look at the screenshot. Are you sure that it IS the standard voltage? My FX chip didn't have that high of voltage.

Yeah, I was just about to say, neither did mine. But AFAIK there isn't a "standard" voltage, the VID will be different from each chip due to binning so the stock voltage will be different. So your advice is sound, set everything back to stock, read voltages, and manually put that voltage in as a starting point.

 

[de.das.dude]

I did not see anything about LLC in my Bios.



I am going to do that when I figure out which size my chips are so I do not buy the wrong heatsinks

Here is a pic of my Motherboard.

http://www.cpusolutions.com/store/pc/catalog/970a-g43_1435_detail.jpg

Do these appear to be the right size copper sinks for those chips?

http://www.enzotechnology.com/images/mos-c1_photo2.jpg

i think this was made for MSI motherboards, and should fit yours. just measure the hole distance in your motherboard. and compare to the specs here
http://www.frozencpu.com/products/7...yte_-_87mm_x_16mm_x_296mm.html?tl=g40c16s1859

or look at these too
http://www.frozencpu.com/cat/l3/g40...t_HeatsinksCoolers-Mosfet_Chipsink-Page1.html

still wont fit?
get these http://www.frozencpu.com/products/7...m_x_65mm_x_12mm_-_10_Pack.html?tl=g40c16s1859 and put them on individually.

 

[newtekie1]

That is what I suggested in his other thread.

 

[de.das.dude]

so did i. i thought it was someone else there :S i suck at remembering names.

 

[terrastrife]

Yeah, I was just about to say, neither did mine. But AFAIK there isn't a "standard" voltage, the VID will be different from each chip due to binning so the stock voltage will be different. So your advice is sound, set everything back to stock, read voltages, and manually put that voltage in as a starting point.

Pretty much every Piledriver is 1.4125v at Turbo, being the highest of the VID range.
I've personally used two 8350's with this VID and a 6100. The Default P1 state has also been 1.375 on all 3 CPUs. Give or take 0.05v depending on the motherboard for the real voltage.

AMD don't really have crazy fluctuating default VIDs like intel, which is why you can undervolt the VID to the extreme very often. I have a II 240 which undervolted by ~.35v, my old 1055T went down .5v. never undervolted the 8350s however because I only ever went upwards in clock speed