Help a newbie Overclock his AMD 8350

[Blacksun]

Hello everyone.
I will be completely honest with you. I am complete noob when overclocking is concerned.
So I turn to you for advice.
When I first built this pc, I ordered a CPU cooler because I had plans to overclock the cpu anyway, (see pc specs in my signature).
Now I think the time has come, to squeeze every ounce of power I can get from this CPU.
I have already read some guides and went ahead and removed the power saving settings in the bios.
I have gathered info from CPU Z, HW and the bios, compiling screenshots which you can find at this link:

https://imgur.com/a/HN10J

Software wise, i already have Prime95 at the ready.
I know that for extreme overclocking liquid cooling is needed. My objective is getting a modest and stable overclock on my modest air cooler.
But I wish to do it carefully, and that is why I ask your help.
Thank you very much in advance.

 

[Blaylock]

Welcome again Blacksun.

For overclocking youre FX 8350 three things will define you're limits.

1. Cooling - The Noctua you have is a good cooler and should not limit you for a moderate OC - GOOD

2. Motherboards Powersection (a.k.a VRM/Mosfet's)-MSI is notoriously bad for overclocking AMD CPU's (especially the 8 cores). At least you DO have one of their better boards so it may limit you, but it's still possible. The MSI 990FX GD65 had an 8+2 Powerphase and it's heatsinked. - OK

3. Silicon Lottery-Most AMD CPU's are fairly good at overclocking, but there are exceptions. This is always a crap shoot and you won't know until you start putting the nails to it so to speak.



You've mentioned you've read a bunch of stickies. That's great that should always be the first step. Get very familiar with them. If you follow Dolks guide for the Phenom's it's more of a step by step guide and it will overclock almost the same as your FX.

Start by opening CPUz, HWMonitor and Prime95(P95) with your rig still at stock. Run P95 in blend mode for about 20 mins watching you're temps. The main ones to watch are Package and socket temp. The package temp is basically the internal temp of the CPU. Keep that under 65c. If it starts to get above that then stop the test. The socket temp is usually called something CPUTIN1 or something similar. You will be able to identify it as your rig gets hot as it will climb in temp along with your package temp. Keep this under 71c, again ending the test once it gets there.

Post your completed results here so we can see how much temperature room you have to play with. To post images click the "Go Advanced" button instead of the Quicl reply then add using the paper clip. Having images hosted here makes it alot easier for us to help you.


EDIT:

Here is an example of what you should include in your screen shot (Complements of RGone).

 

[Blacksun]

Ok here we go

Ok here we go.
The package temp never climbed above 55c°
The socket temp reached the 71°C around 9-10 minutes after test start.
I hope i grabbed the images correctly.

 

[Mr.Scott]

You're all done already if you're that hot at those volts.
Turn those fans up.

 

[bob4933]

Something isn't right if you're maxing temps this early at 1.28v. May have to reset your cooler with some new tim.


Edit: your voltages are jumping around quite a bit. Did you disable c6t and cool n quiet?


Edit2: turbo needs to be turned off as well.

 

[Blaylock]

Well, yes there's good news and bad news. The good news is your CPU cooler is doing an OK job and you have about a 10c margin to work with on the CPU. The bad news is your socket is overheating even at stock. This is very common with MSI unfortunately. My brothers system has the 6 core variant and is doing the exact same thing.

What will you be using this computer for? Gaming, Benching, Mining, Video encoding, E-mail/web surfing?

 

[DarkApollo]

Yikes! Those temps at stock... Honestly, for the $70 you spend on that tower cooler you could have gotten a more efficient AIO, or at least started on a custom loop WC.
You wont be able to move anywhere with those temps.

 

[Blaylock]

That's a different subject, but for the money a top air cooler will generally out perform the aio dollar for dollar. Custom loops are going to be pretty pricy and if he's just gaming not worth the investment IMHO.

 

[Blacksun]

I was quite suprised to see those temps.My system is aimed for gaming, and light video editing/ encoding.
Perhaps those temps are due to low airflow through the case?
The case fans are the stock ones, and I cannot push them further trough the case fan controller, I will try speedfan tomorrow.
I will also recheck the Bios settings, and redo another test.
Thanks for all the replies so far.

 

[DarkApollo]

Check your TIM spread too.

 

[bob4933]

You really should think about reseating your heat sink. Are you 100% positive you applied the TIM correctly, put the right amount on, didnt over torque one side, etc. For something so seemingly simple, it can quite easily be misdone, and you'd never even know until you got the point where you are at now.

 

[Blacksun]

I will check that it is seated properly.
I also have two other aftermarket fans, that I can install, a Noctua NF-a14 and a cooler master mega flow.

 

[DarkApollo]

I was getting crazy heat on my Raystorm and I pulled the block off. The TIM didnt spread correctly and it more like peeled off like clay instead of a paste. Reapplied the TIM, spread it with a credit card and the temps dropped by about 15*! Yeah it was that bad.. The block was actually not even touching the core because the TIM hadnt spread out enough.

 

[Blaylock]

I know it should be a given but watch this quick video on proper TIM application. It doesn't take as much as most people think.

 

[Blacksun]

Good news! (At least i think so)

I Did another test.

Prime 95 reported all "workers" passed the test.
tmt1 newer went above 65C°
Package stayed at 50-51°.
Is there some hope?
(I went into bios and triple checked that power saving features are turned off).

 

[RGone]

Blacksun, let me make a suggestion in how you capture your images since you have hidden some data that is valuable to us trying to help but not at your keyboard or monitor.

Grab the bottom of HWMonitor 'window' and pull it down until we see all voltages and the hard drive temps. And then do n0t cover up the hard drive temps with the CPUz window. For me and I will not say that is for everyone...but I don't care if you even show the P95 Screenie, since you could just say it failed with rounding unexpected or one core failed unexpected but I don't have to see P95 Blend failed, but the information I mentioned as being not shown above is critical 'sluething' information for us not at your location.
RGone...ster.

 

[bob4933]

Your voltages are STILL jumping around mate, you really need to turn off CnQ, Turbo and core 6 state.

 

[Blaylock]

If you haven't changes anything from your imgur pics you voltages changes are likely from still being set to auto. You'll want to change in the BIOS the CPUv to about 1.300-1.325v as that's where you peaked during P95 run. While there ensure the CnQ, C6, spead spectrum etc. are all disabled. I'm not sure what the EuP 2013 is, maybe someone else can define that one.

So you ran an 8 minute P95. Once you make the above adjustments see if it can make a full 20 minutes.

If this rig was used just for gaming there are a few things we could do to "get you by" with your socket temp since no game is going to stress your CPU at 100% load for extended periods like P95. However, video editing will.

 

[RGone]

Blacksun...

...similar to this has already been posted but this has got the what to do's listed in text along with the pics. What should be off / on in bios and what to set windows performance mode to also is listed. If you setup as described below, then the theory is we will no longer see cpu voltage jumping all over the place.

There is one other thing that might have to be investigated and that is if the cpu speed is also bouncing around as the voltage does. But setup as below and we will see what we see.
RGone...ster.

This is what we need to see for sure and a real good starting point.

Normally during setup and testing we disable C1/E, C6, Cool N Quiet, APM, TurboCore and in windows performance manager itself we set to "performance" mode. ALSO if you have HPC in bios you would ENABLE it. That way there are not "other" settings messing with the overclocking process. Some of those settings are not available on all models of cpu but where in evidence we disable for setup of overclock process.

CPU Tab in CPUz from CPUID com

Memory Tab in CPUz from CPUID com

SPD Tab in CPUz from CPUID com

And this is screen capture of HWMonitor (free version) from CPUID com
HWMonitor has been scrolled enough and large enough to show Min/Max of Voltages and includes the CPU CORE TEMPS / "package" temps fully visible. Latest versions of HWMonitor show the CPU Core Temp as " a Package Temp" and is only shown as a single temp since there were never multple, individual core temp sensors anyway.

This capture is made of HWMonitor after it has been open and running on the desktop logging Min/Max temps and voltages while Prime 95 was running Blend Mode test on all cores for at least 20 mins and then the capture of HWMonitor was made and it shows the Min/Max temps and voltages before P95 Blend was started and while running P95 Blend mode and gives much greater insight into how the system is performing without guessing.

In order to attach screenshots of INDIVIDUAL images as suggested, first crop and capture the images with Snipping Tool found in Windows Accessories or equivalent. Then click on Go Advanced, a button at the bottom of every new post window. Then click on the little paperclip tool at the top of the Advanced post window when it opens. Clicking on the paperclip tool brings up the file browser/upload tool and the rest is fairly obvious.

 

[Blacksun]

meh

This time my TMPN1 held for eleven minutes before reaching 71°C
Package temp peaked at 56°
CPU voltage set to 1.302500V ( I Cannot set The CPU voltage through numbers only from a drop down menu).
Windows set to performance mode HPC enabled.
P95 reported zero core failures.