Difficulty getting my FX-6300 OC stable.

Status
Not open for further replies.

tensai27

Honorable
Sep 22, 2013
242
0
10,710
Okay so I am overclocking my FX 6300 and I am shooting for 4.4-4.5ghz. I am using the hyper 212 evo cpu cooler and the asus m5a97 r2.0 motherboard. I am overclocking my cpu to 4.4ghz at the moment and it is requiring an unusually large amount of voltage and before I conclude that I just got a bad chip I was hoping someone could point out if I am doing anything wrong or if there is anything else I could try to get it more stable on less volts.
These are my full specs in case it is relevant:
FX 6300 w/ hyper 212 evo
1gb 7850 overclocked to 1040mhz on stock voltage(1.138v)
7200rpm internal hard drive
Asus M5a97 r2.0
8gb ddr3 1333mhz (4gbx2)
SeaSonic S12II 520w
No optical drive installed
Here are some screenshots of my bios config thus far:
http://imageshack.us/a/img822/3498/yc08.png
http://imageshack.us/a/img18/5234/yp12.png
http://imageshack.us/a/img13/2779/1zh8.png
http://imageshack.us/a/img5/2747/gsfw.png


Thanks in advance and any help/insight will be greatly appreciated.
-edit: just noticed the ht link speed on two of my images conflicted but I have ht link currently set to 2200
 
Solution
Just checking back.

Glad you're getting closer to your goal. I can't blame you for being so cautious, especially considering this is new to you. :)

All the voltages in my OP are still below the max safe limit by a bit. So you should be OK as long as you don't exceed those and as long as you have good case ventilation/cooling.

I generally stop at 56c load temp or 1.475v, whichever comes first. That is well below the limit for this CPU, but I like to keep some longevity built in my overclocks. 😉
Well to be honest I had to overvolt by an entire .1 Volts just to get mine stable at 4.2 GHz. You might try playing with load line calibration, but I chose to keep my load line calibration at the default settings due to an interesting article I read about it, which I unfortunately can't find anymore. However, the gist of it was that messing around with load line calibration could actually cause you to temporarily cause too much of an overvolt and actually produce more heat than you would by just increasing voltage manually. The other thing I found is that it's a bitch to get an FX-6300 stable if you're playing around with the BLCK, but it's so much simpler if you just stick to the multiplier and for RAM just sticking to memory profiles.
 

Thanks for reply! As for load line calibration, it's to my understanding that it increases stability when overclocking at the expense of extra heat. I left it enabled since my temps where low enough but I am wondering now if I should try disabling it. Is there really any risk of damaging my cpu by leaving it enabled provided that my temps are fine when stressing? I left ram profiles as is and am oc'ing through the multiplier since it is an unlocked cpu after all. Isn't 1.48v ridiculous though for just 4.4ghz?
 
Ah well I found the article I was looking for, here:
http://forum.overclock3d.net/showthread.php?t=46502

Anyway, I'm looking at your settings. If you're not playing around with LLC, then I suppose it's fine that it requires a .125 V increase to get to 4.4 GHz stable since I needed to go up by .1 V just to get to 4.2 GHz stable and I also didn't play around with LLC. So I don't think any of your parts are defective, I think we can just expect that based on the fact that we're dealing with budget motherboards. So if that's the case, I think you're fine where you are but I personally wouldn't start going any higher than that.
 

I was able to hit 4.1ghz on stock. Sorry I should have mentioned that. I will try changing the ht and NB.
 

Well, I currently have CPU LLC enabled and left cpu/nb at default(auto). Should I change any of those?
 


You can hit 4.1GHz, but how stable is it? We aren't talking about how high you can go, we need to know what is the max stable overclock with stock voltage.

So, what is you max stable overclock with stock voltage?
 



Did a 2 hour occt test. Should I do it for longer?
 


2 hours are good indicator but I would run it a little more. I did mine for 9 hours.
 

I will try another test for longer and let you know how that goes.
 
Alright in the middle of stress testing again at 4.1ghz using occt for 4 hours. 64 bit test with a large data set. Left voltages at stock, disabled turbo boost, changed nb and ht to 2000 and disabled power saving features such as cool n quiet. If the test fails I will have to reassess the max stable clock at stock. I will post back with updates as I go.
 
Alright I'm back. After lot's of stability testing it seems that my max stable clock at stock is a lot lower, 3.8ghz. Anything else I should do before I move on and start slowly raising the voltage and clocks? How about ram timings? Also, after I reach my max stable clock at the max voltage I am comfortable with, should I then proceed to raising the NB and HT and then stress? Would that benefit me performance wise?
 
Just to clarify, I'm stress testing 4 hours on occt. max stable clock at stock voltage(1.36v) was 3.8ghz so I am now testing 3.9ghz at 1.37v. If it passes then I will raise the clock up to 4ghz @1.37v and so on until it fails then raise the voltage one and try again. Nothing else I should be doing
 
Update: 3.9ghz passed at 1.37v now testing 4ghz but it seems to be failing a lot. Up to 1.4v already as it has failed at everything below that. Did I hit the wall already? It's not even my turbo speed yet:??: I'm comfortable pushing it up to 1.45v-ish provided the oc is decent.
 

So at 4.1ghz I should raise the NB to ~2200mhz? Also, sorry but what do you mean by ADM?
 
Status
Not open for further replies.