SOLVED AMD FX-6300 Overclocking

[rescuetoaster]

I would just roll with it and see what happens - that case isn't doing you any favors, which I'm sure you know. Maybe at some point in the future you can invest in a newer and better case, since you can always re-use the case years down the road. Rosewill often has sales on some nice cases, so you could just watch for them in the future.

 

[RJARRRPCGP]

The green options are the cause of the core frequencies dancing around!

I would also disable APM to be sure, too.

 

[Ghost P.]

The fans helped a lot. Upped to 4.5ghz. 4.6ghz is impossible for me without hitting 60+/- degrees, it requires 1.43v.

I fell asleep while the test was running, originally I intended it to be for around 1 hour, lol.

And why should I disable APM?

 

[Johan45]

Yes that did help a lot, see what a bit of fresh air can do. For you Ghost since that's a Giga board I would leave APM enabled, just make sure you have HPC enabled also. When you're done you OC set the CBP ratio to the same multi you use for your OC and that will eliminate the odd double boot at start up.

 

[Ghost P.]

Yeah, I have both APM and HPC enabled. So, I should set CBP to auto and adjust the CBP multiplier to the CPU clock ratio to eliminate the double boot?

Last time I did that I didn't check if the system double booted on start-up, but my cpu speed in the task manager would bump arond all the time and didn't stay constantly at 4.4, so I disabled turbo boost(CBP).

 

[Johan45]

You may need to fiddle with it a bit ghost but that's what will stop the double boot AFIK , I don't have that board so I can't test it out. Maybe try enable instead of auto for the CPB and use the same multi you're using for your OC clock.

 

[Ghost P.]

I updated my bios to the latest beta version and the double boot was fixed without enabling turbo boost.

 

[RGone]

I updated my bios to the latest beta version and the double boot was fixed without enabling turbo boost.

Good that the new bios fixed it. Different brands and sometimes even different model of boards do different things when setup with what seems the 'same' settings. This can happen because of somewhat different physical differnces on how each board is constructed.

Since it is better to 'know' that TurboCore is truly off and not trying to do something in the background it is truly better to have it in a known off condition. Good deal.

Your temps and voltages would indicate you could perhaps get a little more 24/7 useful speed out of your rig. I will say this though after many hours of benches and real world use, it takes at least a good solid 200Mhz extra cpu speed to really know or feel a difference. A bench might indicate a slightly higher score but for me real world feel of any perceived increase only comes at at least 200Mhz more on the cpu. So when i cannot get the full 200Mhz or so increase, I usually settle for the lesser speed since it runs cooler and most of the time quieter since the fans don't have to be run quite so high.
RGone...ster.

 

[Ghost P.]

Your temps and voltages would indicate you could perhaps get a little more 24/7 useful speed out of your rig.

There is a huge temperature and voltage jump in order to reach 4.6ghz.

Max temp was 61C and average was around 58C, at 220mhz ; 21X multiplier with voltage set at 1.44v and LLC at medium.

 

[Johan45]

That's still within acceptable levels Ghost but you be the judge it's your parts.

 

[Ghost P.]

So you suggest that running at temperature between 60-65 is still in the safe zone and won't damage the CPU? Again, this is the first time I'm using an AMD cpu and I've read somewhere that the max safe temp for the FX-6300 is 62C and TjMax is 70C.

Should I use Core Temp or CPU-ID(Hardware monitor) while stress testing?

 

[Johan45]

Between 60 and 65 is fine in my books Ghost. Just continue what you're doing and keep an eye on the socket temps. As you increase voltage they will also rise keep them below 75c.

 

[Ghost P.]

Socket temp usually stays below CPU Core temp, at least in the stress test that I have done before. I'm worrying If I screw up anything by setting the voltage too high though, right now I'm trying to stabilize the cpu at 4.6Ghz with 1.443v and it seems like i'll have to increase to 1.45v. Prime95 failed after 1 hour and 3 minutes, I'll try again the same stress test tomorrow after I boot up the PC.

Temperature was between 58C and 62C.

 

[Johan45]

That all sounds good Ghost. Most likely your board won't let you set a voltage that would do permanent damage to the CPU. AMD has said up to 1.55v for air/water is OK

 

[Ghost P.]

4.6Ghz
220mhz, 21X, 1.45v, LLC - medium

1h30min stress test

I finally got it stable, but temps were meh. The highest I noticed was 66, but it jumped to 68 because I opened Google Chrome. It usually stayed around 62C-64C.

Next I'm going to do 225x20.50 to see if there is a difference between temperatures.

 

[Mandrake4565]

Looks good, temps are a bit warm for my liking but you will most likely never see similar temps during 24/7 usage

 

[Ghost P.]

No difference in temps. I'm really not sure about going higher than this. I don't think there'd be a difference If I turned off HPC/APM either.

 

[Johan45]

Well I wouldn't go any higher, you're definitely pushing the limit there Ghost

 

[Ghost P.]

Yeah. On a second topic, I opened the case today to clean the fans and decided to connect them directly to the PSU instead of the mobo. I used one PSU 4-pin to connect both of them. Should've I connected each of the fans in a different 4-pin connector?

 

[RGone]

Heck no man...one 4 pin Molex carry two fans pretty darn easy.
RGone...ster.