FX6300, M5A78L-M LX and AMD Turbo Core Crashes

[RoyMust4ng]

Hello guys! I've read this forums a lot but never asked or answered anything until today.

The short story is I bought a VGA and a processor a few months ago. I got a FX6300 and updated my mobo's BIOS to be ready (M5A78L-M LX Bios 1603). After a month of normal use I started getting hardware related BSOD's (101 and 124). After 3 months of tracking, enabling and testing everything i got to pin point the issue generating the inestability: AMD Turbo Core Technology. So i have it disabled to avoid crashes and it worked.

Speccy link with my full specs: http://speccy.piriform.com/results/HduqvTqBXYa4u1iKex1kINV

I tried to reach ASUS support but it's just terrible, they answer like some call-center kid who know absolutly nothing. Reached AMD support forums as well, no answer on the matter after a week. I did find some info here, another user had the same issue, not the same specs, but he had the symptoms. After reading the thread i found out some info on the issue. According to the users in that thread the origin of the problem could posibly be the mobo not being able to mantain the Vdrop or some stuff like that. And the suggested solution was to change the LLC on his motherboard. LINK: http://www.tomshardware.com/answers/id-2279689/error-source-processor-amd-9590-8core-freezes-system-lot.html

And here comes the question itself: how do i do that LLC change? I know how to get to the option in the BIOS (it wasnt really hard since i've investigated that bios several times in these months 😛) but i dont know how to change it.

The setting is now in AUTO, the side screen info says this:

-Step: 3.225%
-0: 0%
-100: 100%
-Auto: 48.375%

And when I try to change the setting i get a list of all the posible values between 0 and 100 at 3.225% steps.

The question is: should I raise it? Lower it? How much in either case? Should i just leave Turbo off and live my live happy like this?

Thanks if you took the time to read all of this 😉

Roy.

 

[GhislainG]

You may want to read http://linustechtips.com/main/topic/24019-load-line-calibration-why-overclockers-should-care/ and other similar articles before increasing it.

 

[RoyMust4ng]

Thanks, that was a really great read.

Roy.

 

[naturesninja]

You can go through all the changes you want, and it won't change the fact that your mobo only has a 3+1 vrm setup.

 

[naturesninja]

Try turning off turbo, overclocking to 3.8 GHz @ 1.2 volts. You can then move from there to find the highest stable clock with the lowest voltage possible. The overclock I recommended is a good starting point for your motherboard. You will likely not see a good stable overclock at anything higher than 4.0GHz with your motherboard even if you have a top down cooler blowing air right onto your VRMs.

 

[RoyMust4ng]

I didnt understand the 3+1 VRM setup you said, what's that? Imnot into OC, I know barely nothing about it and dont really plan to do it any time soon. The thing i want, if it is possible, is to enable that AMD Turbo option to have my processor to performe as good as possible with out getting into OC.

Thanks for your answer 😉

Roy.

 

[GhislainG]

I didnt understand the 3+1 VRM setup you said, what's that? Imnot into OC, I know barely nothing about it and dont really plan to do it any time soon. The thing i want, if it is possible, is to enable that AMD Turbo option to have my processor to performe as good as possible with out getting into OC.

Thanks for your answer 😉

Roy.

A 3+1 phase VRM probably is too weak to supply the correct voltage when turbo is enabled.

 

[naturesninja]

I'm sorry, I should have elaborated. Like GhislainG said the VRMs likely can't keep up with the abrupt voltage change when turbo kicks on. The VRMs (voltage regulator modules) feed power to your CPU when needed, and the vishera chips (even 95w) don't play well with low count VRMs.
That's why I recommend trying to overclock to get as close to the 4.1GHz turbo would give you. Trying it this way should be more stable than allowing turbo to turn on, since you won't incur sudden voltage changes. The 3+1 VRM's should allow you to use around 1.2-1.3v stable. The chip requires a little over 1.4v for turbo, but these chips can be overclocked to over 4.5GHz using that voltage and the proper motherboard. So, if you can get yourself to 4.0GHz or so on 1.2-1.3v, don't think of it so much as overclocking. Think of it more like turbo's always on and you are "undervolting."

One last thing to check is that you DON'T have cool and quiet features turned on in your BIOS as this may not allow for enough voltage to be supplied to your CPU when it asks for it. If you are still having problems, I can only recommend getting a top down cooler for your cpu to help blow air onto your VRMs and keep them cooler to allow a stronger supply of voltage without overheating. Or get yourself into a new motherboard. Also, what PSU are you using? I couldn't find it in your link.

 

[RoyMust4ng]

Well, here i am with the last news and the final solution to this issue.

As naturesninja said, it is a mobo limitation. After reading about LLC and some other stuff, i gave it a try. Raised it one step and the crash happened again. I put it on one more step and tried to force the crash but it never happened. Increasing the LLC worked like a charm. The bad news is that my proc was melting 😛

I started to check the temperatures (as recommended in everywhere 😛) and it looked good until I put some heavy load on it. Playing FIFA 15 put the core temp (logged by HWiNFO64) around 67ºC with peaks on 73ºC. I decided to get a heavier game so i played Shadow of Mordor and there i had it worse. Peaks over 80ºC and smaller peaks around 75ºC. Since imnot gonna get any aftermarket cooling system, then this is the end of the road for me.

With the Turbo off i dont get any kind of crashes and my games run just great, i just wanted to use all the features on the proc i bought.

So thank you everyone for reading and posting 😉

Roy.