fx-8370 thermal throttling at ~40.5°C

[Bluefalcon13]

Not really... LLC just prevents droop. Whether you hit 1.5v with or without llc is still 1.5v.

My board says otherwise

(Before my H80 died: idle - 1.416vcore, full load - 1.44vcore. LLC set to high (medium results in vdroop)

 

[bassnut]

Remember the VCore is dependent on your cooling. As good as the 212 is it still probably falls short of you H80 thus would explain your higher VCore. EarthDog is right. 1.5v is nothing more then 1.5v LLC just adds a bit to offset your VDroop.

 

[Bluefalcon13]

Gonna have to say that is incorrect, based on stuff I will PM to both of y'all. Not trying to thread-dump.

 

[Bluefalcon13]

I think we are all roughly talking about the same thing, but vdroop is a function that involves current, which is why it can appear to 'bump' your vcore voltage. Either way, the more granularly you can control LLC, the better you can mitigate this. In example, if I require 1.416v vcore for stability at 4.5ghz, medium LLC may result in vdroop, which results in instability. High may result in more voltage getting to the CPU than absolutely necessary, increasing temps, and therefore limiting overclocking due to thermal solutions.

 

[EarthDog]

In the end, if you need x.xx volts to get stable, it will be stable there regardless of how you get the voltage and regardless of droop or llc. LLC simply mitigates droop and sometimes causes vraise (what I call it when LLC goes over what is shown at idle in windows). If it goes over and you don't have granular controls over llc, lower vcore which always has granular control.

That said, I see what you were trying to say earlier in the post above.

I digress.

 

[RJARRRPCGP]

Hello.

I have a mobo gigabyte 970a-ds3p

The DS3P is even lower-end than the D3P! I'm not surprised that the DS3P is giving you issues.

And the D3P can work fine, believe it or not, but seems to be hit and miss, so I wouldn't get anything less than the UD3P.

And I would disable Cool n' Quiet, APM and C-states (IIRC).

And enable HPC.

 

[roberto rossi]

I think I will get the 970a-Ud3p,,,,,,,, revision 1.0 is said to be 8+2, but revision 2.0 just does not inform a think about it....... Can I trust revision 2.0 from gigabyte 970a-ud3p?

By the way........ when I was cooling with air cooler, temps without overclocking came to 65°C (70°C when benchmarking)!!!!!! kkk then I bought a simple H55 water cooler, and temps does not overpass 45°C even benchmarking (w/0 overclocking of course)....... if I get a ud3p, apart from my ds3p, I hope it has a better heat dissipation, not just the 8+2 stuff...... dont u agree?

 

[RJARRRPCGP]

I think I will get the 970a-Ud3p,,,,,,,, revision 1.0 is said to be 8+2, but revision 2.0 just does not inform a think about it....... Can I trust revision 2.0 from gigabyte 970a-ud3p?

By the way........ when I was cooling with air cooler, temps without overclocking came to 65°C (70°C when benchmarking)!!!!!! kkk then I bought a simple H55 water cooler, and temps does not overpass 45°C even benchmarking (w/0 overclocking of course)....... if I get a ud3p, apart from my ds3p, I hope it has a better heat dissipation, not just the 8+2 stuff...... dont u agree?

Yes!

 

[Mandrake4565]

I think I will get the 970a-Ud3p,,,,,,,, revision 1.0 is said to be 8+2, but revision 2.0 just does not inform a think about it....... Can I trust revision 2.0 from gigabyte 970a-ud3p?

By the way........ when I was cooling with air cooler, temps without overclocking came to 65°C (70°C when benchmarking)!!!!!! kkk then I bought a simple H55 water cooler, and temps does not overpass 45°C even benchmarking (w/0 overclocking of course)....... if I get a ud3p, apart from my ds3p, I hope it has a better heat dissipation, not just the 8+2 stuff...... dont u agree?

Yes having a 8+2 Vrm section will help. About the Giga board itself on paper they are good. They've had many revisions though and I'm not 100% up on which one is good and which is bad.

 

[roberto rossi]

Hi guys........ I've bought my 8+2 motherboard, gigabyte 970a-u3dp. My amd cpu fx-8370 is not throttling anymore, even when the temperatures reach more than 60°C (both CPU and MB TMPIN1) ... when I use HeavyLoad. I've noticed that, even with this board, the mb temperature TMPin1, when hot, takes time to get cold again.... instead of the cpu, that quickly lowers the temperature...

Why does heavyload raises more the cpu temperature than prime95? Both use 100% cpu,,,

I've noticed that my cpu throttles now when the temperature of one of the sensors in my motherboard comes up to 70-71°C when performing a torture test. This is pretty understandable,. thanks

 

[bassnut]

I'm not familiar with the gigabyte sensor outputs so am not certain that Tempin1 is your socket temp. If you are getting throttling 70 / 71 then something important is associated with that sensor. If you have easy access to your case and can watch your temps while using some compressed air to gently blow on your heat sink for your mosfets and you NB, seeing a temp drop will indicate what that sensor is for. My guess it is one of those two but then that is just a guess.

Be carefully not to hold the can upside down doing this as it will spray your board with liquefied oxygen. Not sure but willing to bet this would be bad.