Having a hard time overclocking my cpu(fx-6300)

[wuffll]

So, as the title says, I am trying to overclock the fx-6300. I can get it somewhat stable at 3.6 ghz but anything beyond that I can't(isn't stable in prime95 during a 6h test[blend]). When I try going over 3.6 it starts out good at the beggining but after some time it goes down to 24 mhz(at 4ghz). I have a m5a97 r2.0 motherboard. At stock speeds, idle temp is around 27c(probably +5c becasue of innacuracy) and while under load its around 47c.

 

[CountMike]

Can you describe how you are overclocking ?
Do you have AMD Overdrive program ?
Which CPU cooler are you using ?

 

[wuffll]

I keep changing the multiplier until my pc freezes/shows an error. When the pc freezes I change the voltage while trying to maintain the temp under 62c. I keep turning up the voltage (and the multiplier) until I get to temps around 60c.
I do have AMD Overdrive but I rather use the BIOS.
I have a LC-CC-95 cooler.

P.S. Not trying crazy overclocks, just trying to get it around 4 ghz.

 

[omgazombie2k]

try changing your bclk along with your multiplier, my pc hits 4ghz with an 8320 while using only 1.25v compared to the normal 1.425v it wanted to use before and it's stable as heck

 

[wuffll]

Increasing the bclk(multiplier) crashes my computer within 2-10 mins(during prime95). If I increase the voltage it gets into very high temps(62+C) after 2-4 hrs of prime95.

 

[oobymach]

Honestly hours of prime 95 is overrated, I try the small fft's test for about 5 mins to make sure it doesn't freeze and the default test for 4-6 minutes until it passes the first hurdle without any cores failing. I'm at 4.4ghz but with 210fsb and 2730ht link.

I had to overvolt the cpu and the cpu-nb a bit to get it stable but unless you plan on pushing your cpu for hours on end in tasks that demand accuracy like medical computing or something like that go ahead and push it until you get 4.2-4.6ghz and do quick tests for stability rather than long hours.

The small fft test will heat up your cpu and cause the system to freeze within a couple mins if it's unstable, the regular p95 test will fail workers if the voltage is too low/high or your ram oc is off, or if you oc the nb without overvolting.

Good luck, I've spent alot of time testing my own cpu so hopefully my advice helps. When pushing my oc beyond 4.4ghz the p95 tests both fail under stress within the first couple mins so that's your red line, pushing it to the red will cause glitches in games and such so when it starts failing dial back the cpu oc a bit.

 

[wuffll]

@oobymach What method should I use to see the stability of my cpu? Lots of people like hours of prime95 because it makes your cpu struggle quite a bit, which is the point, to test out the stability. I'm just wondering if I am doing something wrong in the BIOS.

 

[oobymach]

Every board overclocks differently, you just have to keep trying until you find a sweet spot, I'm now at 2795 ht and 215fsb @ the same 4.4ghz cpu and games seem to be stable and very fluid, I was up to 220fsb but kept either dropping cores in p95 with default setting before 6 min mark or would cause freeze up with p95 small fft test.

The small fft test will heat your cpu up and test strictly the cpu overclock (and is likely to fail in the first 5-6mins by freezing the computer so if you can get to 7 mins or more you're fine to keep going) where the blend test stresses everything so if there is a weakness in any of your ram, cpu, northbridge, clocks or voltages the blend test will fail within the first 6 mins.

Even if it fails after this point I would still call the overclock stable because almost nothing is going to stress your cpu to that point just gaming on it, but if you're looking for other testing tools OCCT is a good one to fail quickly if your oc is too high but again it's extreme stress that is likely to cause an issue and you're unlikely to cause your cpu extreme stress under normal circumstances. I fail OCCT at just over 1min30sec.