[SOLVED] Getting nowhere on my overclock, really frustrated

[Bobbrugge]

I am tired of asking over and over again for help and getting nowhere, so this is my last time.

I have
Asus 970 pro gaming/aura

Amd 6350

8 GB ddr3 Hyper X savage x 2

Hyper evo 212 cooler

1500W psu

Raedon r290x GPU

I have read the overclock tutorial, read everything I can find, watched videos for the board, and the cpu. And still I am getting just no where.

The issue lies with my thermal margin and my cpu voltage.

So first, I ran prime 95 on small fft’s for 15 minutes with everything set to optimized defaults in the BIOS to get a baseline. Coretemp shows VID at 1.3875 while hwinfo shows cpuVID at 1.388.

BIOS shows CPU manual voltage at 1.38, the monitor tab on my BIOS shows 1.38. When I set the cpu manual voltage to 1.38, I reach my thermal margin in 2.5 minutes flat. The tutorial says don’t go any closer than 10C before tjmax, I stopped at 15C before.

With everything set at optimized defaults I go no higher than 34c before tjMax.

So what am I doing wrong then?

If no one will help me make sense of this I will try a different forum, or just by a Dell and call it good.

I have been trying to overclock this thing for years and now with everything shut down, I have time to get good and mad at my lack of production here.

Can someone with real experience and a little time please help?

 

[CountMike]

I am tired of asking over and over again for help and getting nowhere, so this is my last time.

I have
Asus 970 pro gaming/aura

Amd 6350

8 GB ddr3 Hyper X savage x 2

Hyper evo 212 cooler

1500W psu

Raedon r290x GPU

I have read the overclock tutorial, read everything I can find, watched videos for the board, and the cpu. And still I am getting just no where.

The issue lies with my thermal margin and my cpu voltage.

So first, I ran prime 95 on small fft’s for 15 minutes with everything set to optimized defaults in the BIOS to get a baseline. Coretemp shows VID at 1.3875 while hwinfo shows cpuVID at 1.388.

BIOS shows CPU manual voltage at 1.38, the monitor tab on my BIOS shows 1.38. When I set the cpu manual voltage to 1.38, I reach my thermal margin in 2.5 minutes flat. The tutorial says don’t go any closer than 10C before tjmax, I stopped at 15C before.

With everything set at optimized defaults I go no higher than 34c before tjMax.

So what am I doing wrong then?

If no one will help me make sense of this I will try a different forum, or just by a Dell and call it good.

I have been trying to overclock this thing for years and now with everything shut down, I have time to get good and mad at my lack of production here.

Can someone with real experience and a little time please help?

I've read it few times but didn't see which frequency you OC-ed to nor I can understand if you are doing it thru BIOS or AOD ?
Are you changing multiplier or/and FSB to overclock.

 

[Bobbrugge]

I've read it few times but didn't see which frequency you OC-ed to nor I can understand if you are doing it thru BIOS or AOD ?
Are you changing multiplier or/and FSB to overclock.

Thanks for responding.
I found if I change my llc to medium instead of high or extreme, I can suddenly and amazingly shed heat.
It has taken me nearly ten years of screwing with it off and on to find that ONE setting.
I am overclocking through BIOS. currently set at 4.5Ghz with about 30C thermal head room left before I melt it. which is amazing since before I found the medium llc setting, I couldn't run stock numbers manually set without thermal issues.
I am changing the multiplier currently as I have just started this process.
The only other real issue I am having is with HWinfo and CoreTemp not reporting correct voltages, or I have them set up wrong. Currently they show 1.38 +/- .005 for CPU VID no matter what the actual voltage is set at in BIOS. even the monitor tab reports less voltage than the 1.425 V I have it set at.

I am trying to hit a safe thermal boundary through monitoring at 100% load, then intend to bring my multiplier up more until it gets crashie. I have read several threads and tutorials on here that recommend baby steps to avoid letting out the magic smoke in the computer. Since I can't seem to get any response out of 2 different programs on CPU V, I am unsure what else to do to have this happen as safely as possible.

If it matters, I am upgrading to windows 10 right now.

 

[CountMike]

Thanks for responding.
I found if I change my llc to medium instead of high or extreme, I can suddenly and amazingly shed heat.
It has taken me nearly ten years of screwing with it off and on to find that ONE setting.
I am overclocking through BIOS. currently set at 4.5Ghz with about 30C thermal head room left before I melt it. which is amazing since before I found the medium llc setting, I couldn't run stock numbers manually set without thermal issues.
I am changing the multiplier currently as I have just started this process.
The only other real issue I am having is with HWinfo and CoreTemp not reporting correct voltages, or I have them set up wrong. Currently they show 1.38 +/- .005 for CPU VID no matter what the actual voltage is set at in BIOS. even the monitor tab reports less voltage than the 1.425 V I have it set at.

I am trying to hit a safe thermal boundary through monitoring at 100% load, then intend to bring my multiplier up more until it gets crashie. I have read several threads and tutorials on here that recommend baby steps to avoid letting out the magic smoke in the computer. Since I can't seem to get any response out of 2 different programs on CPU V, I am unsure what else to do to have this happen as safely as possible.

If it matters, I am upgrading to windows 10 right now.

Trust AOD but only for monitoring, Even 5c margin is OK, TjMax reported for FX processors is not shutdown temperature like for many others but only slowdown safety.
I have my FX6350 on GA-990XA-UD3 and under Scythe Mugen 3 running at 4.9GHz @1.495v for couple of years now, no problems, Voltage set at 1.45 and LLc3.

 

[Bobbrugge]

so 90c is the throttle temp?
What temps are you running?
On air or water?
I hit 4.5G today @ 72C, going to run stability test over night.
Running 1.49V.
Any tutorials for msi Afterburner 6.1? I think thats the one I have, all the tutorials I can find so far are for 4.5

 

[CountMike]

so 90c is the throttle temp?
What temps are you running?
On air or water?
I hit 4.5G today @ 72C, going to run stability test over night.
Running 1.49V.
Any tutorials for msi Afterburner 6.1? I think thats the one I have, all the tutorials I can find so far are for 4.5

Afterburner is useless for Asus MBs.

 

[Bobbrugge]

Afterburner is useless for Asus MBs.

How does one overclock an MSI GPU on an asus board then?

 

[Karadjgne]

You do know the difference between cpu VID and cpu vcore? VID is the voltage that the cpu demands from the motherboard VRM's. Vcore is what the cpu actually uses. Ideally you'll want the two of them to be @ 0.05v apart at best. Then use positive or negative offsets to offset the VID to just above vcore.

Don't get the two confused or mixed, they aren't the same thing.

Start out at idle. Use cpuz or similar to get vcore. Then set vid to match, using the offsets. Then apply a load and check for stability. Sometimes lowering VID to match results in lowered vcore at idle, so it's a game of adjustments to dial them in.

LLC is a voltage adjustment that works both ways. When you setup the vcore etc that's at idle. The load will drop vcore. The changes in load voltages are called vdroop, which changes almost instantly. The problem is the cpu demand can happen faster than the VRM's can adjust, so LLC is a pre-emptive voltage applied to take up the slack, stability. But it works on both the negative and positive sides, so applying high/extreme llc adds a ton of voltage on top of the vcore, and instead of 1.38 the cpu has constant power spikes of 1.45v (give or take) and thats nothing but pure heat output.

Far better to run 50-70% LLC at best, much lower heat output from the cpu. You only need enough LLC to maintain stability, any extra is waste.