Overclock FX 6300 - Underclock in software monitors?

Soulage

Reputable
Apr 14, 2015
177
0
4,710
Hello all,

I've been reading and reading about overclocking the last couple of days and today was the day I decided to go for a light ish overclock of my FX 6300. I started by disabling the standard things I'd seen in various guides, at least the options my bios would allow ( 970A UD3P - Rev 1.0 - F1 ). Then I changed the frequencies and voltages to the exact values that they were on for auto to give myself that base. After that I set the Multiplier from 17.5 to 20 and the FSB to 210 bringing the clock to 4.2. Saved, exit and ran Prime 95 blend test for around 40 minutes, passed IBT at standard, and ran the Heaven benchmark, all without error. Also logged into Guild Wars 2 and killed some stuff with players around as a quick check as this is the main reason for the performance bump.

My goal is the 4.2 mark anyway due to my cooling, not stock but not great either ( Cooler Master Blizzard T2 ) so I think I have achieved that. However while monitoring temps etc in HWInfo I noticed the chip was being downclocked to 3144.4MHz and the voltages are also alot lower than I've seen posted aswell as CPU-Z having a different value to HWInfo.

This is the first computer I've owned that really has the ability to OC and stuff so I've not got the expierinece before today. My main questions are have I got the correct voltages and is the downclocking a cause of it being too low at what I assume is stock if the values were just copied over? Screen shot attached.

kAKS4lK.jpg


Many thanks for any insight that can be provided :)

Edit: Since posting this I've been back into the bios and it seems the voltage I set is now at 1.2750 instead, yet HWInfo and Core Temp still show it running at 1.1750 with CPU-Z being more accurate at 1.260.
 

Zerk2012

Titan
Ambassador
The answer is not in the information in the first pick you have fanspeed covering tne highest temp you were reading.
That processor has a TJmax of 82C it will then throttle down to cool off go back up rinse and repeat.
Run it again and monitor your temps I belive that will be the problem.
 

Soulage

Reputable
Apr 14, 2015
177
0
4,710
The highest temperature would have been the ones in Core Temp with a margin of 1-2 degrees, being at full load they were stable. I doubt it's thermal throttling causing it. I would have stopped the stress test if it got much over 62. Not meaning to be rude here as I know you're only trying to help but I didn't mention temps being a problem, I'm more concernced about improper voltage. I didn't push to far knowing my cooling solution isn't ideal.
 

Soulage

Reputable
Apr 14, 2015
177
0
4,710
Just an update for anyone that comes across the thread.

After much more reading and monitoring things, it seems it was a couple of things going on.
HWInfo and Core Temp didn't like that I'd set manual values in bios and were showing me some wrong info, such as the core voltage. They did work fine however with optimised defaults set.
After checking the voltage in bios and seeing that CPU-Z was accurate I downloaded HW Monitor made by the same people and ran things again.

It seems what was happening was vdrop/vdroop at the high stress levels of the burner programs so the voltages couldn't be maintained thus the downclock.
It also looked like the voltages were jumping around a bit because of LLC, and Gigabyte didn't make it easy to find the values for each setting as setting them it bios didn't show any value change. The manual has a bunch of descriptions when it comes to other sections to change but when it came to the advanced voltage section it just says "This sub-menu allows you to set CPU, chipset and memory voltages" (duh really?). Anyways I found some info from other people testing LLC settings and set accordingly.

Since everything seems to be in order now, I'd just like to thank Zerk even though it wasn't thermal throttling his mention of throttling guided me to change my search parameters while looking for info.