It's stable but......

I've been tinkering ever since installing my corsair H50 with 2 push/pull 120mm fans. Cooling at this point is not an issue. I've rarely seen 50C since then. I have been able to get 3.8Ghz stable by bumping th CPU-NB to x9 and voltage 1.25 CPU 1.45v. Here's the $100,000 question. I can boot into windows and stable @ 3.7Ghz with NB x8 1.175v and cpu 1.40v I intend to run pretty much 24/7 although likely shut down at night. Is a 100Mhz worth the extra voltage? I mean wear and tear wise? There appears to be a big increase in performance between the 2. I haven't benched it yet but I did do some video conversion and it goes a LOT faster. Opinions? Not sure I'll try going any farther. It wouldn't be useful 24/7 and that's pretty much all I'm really interested in.
Temps
idle 39Cish
load 50C max
If I turn the fans up idle temps will come down but noisy.

Its up to you. Its certainly not going to hurt the CPU. You will get another before you kill it at those volts and temps.

Considering 1.425v is the top of the "AMD safe" range for vCore, and the fact that you're running pretty cool, I wouldn't be afraid to run that 24/7 for several years. (And I mean that literally. )

The cpuNB voltage isn't even out of the "AMD safe" range (1.15-1.30v) so you don't have an issue there at all.

The voltages are probably safe with good cooling but I highly doubt that the performance diff between 3.7 and 3.8 Ghz. is even noticable let alone a "lot". You might want to run some relevant benches to be sure but I've not seen any significant system performance increase for a 100 MHz. change.

beenthere said:

The voltages are probably safe with good cooling but I highly doubt that the performance diff between 3.7 and 3.8 Ghz. is even noticable let alone a "lot". You might want to run some relevant benches to be sure but I've not seen any significant system performance increase for a 100 MHz. change.

Click to expand...

+1. That extra 0.1 ghz amounts to less than a 3% increase in processing speed. Do the math. Divide 3.8 by 3.7. Video conversion time difference could have been due to some background process running during one of the tests that was stealing CPU time.

beenthere said:

The voltages are probably safe with good cooling but I highly doubt that the performance diff between 3.7 and 3.8 Ghz. is even noticable let alone a "lot". You might want to run some relevant benches to be sure but I've not seen any significant system performance increase for a 100 MHz. change.

Click to expand...

Thanks for the input. I'm thinking that maybe I should turn it down to 3.7Hhz but leave the NB alone. Actually I MIGHT be able to reduce voltage to the NB for lowering the cpu clock. There's no issue though so I'll leave it alone. I don't like Prime95. I'll probably get it from the community but I think it represents nothing. Their is virtually 0 chance that I'll ever run into a situation that stresses all of the cores to their max while running multiple threads each. Such a program would be unpopular for obvious reasons. Doing video works them at 65-75% and I'm usually doing other things like playing games while it's going on with no problems. That is what I call a real world test.

Rigit said:

Thanks for the input. I'm thinking that maybe I should turn it down to 3.7Hhz but leave the NB alone. Actually I MIGHT be able to reduce voltage to the NB for lowering the cpu clock. There's no issue though so I'll leave it alone. I don't like Prime95. I'll probably get it from the community but I think it represents nothing. Their is virtually 0 chance that I'll ever run into a situation that stresses all of the cores to their max while running multiple threads each. Such a program would be unpopular for obvious reasons. Doing video works them at 65-75% and I'm usually doing other things like playing games while it's going on with no problems. That is what I call a real world test.

Click to expand...

There's more involved than just the processor load, it's a matter of producing the heat as well. If you keep your systems well maintained then you may have a point since your cooling capacity won't diminish too much over time. If you're like most people, however, your cooling is at its peak right now - or rather a few days ago when you installed the H50 - and it will get worse every hour you run your machine. Dust on the leading edges of fan blades, dust in the rad fins, etc., all will contribute to a decrease in cooling capacity.

If you're still using your computer two years from now, and you're not avid about cleaning the insides, then P95 may be simulating the conditions you'll have then when only stressing the CPU that 75%.

QuietIce said:

There's more involved than just the processor load, it's a matter of producing the heat as well. If you keep your systems well maintained then you may have a point since your cooling capacity won't diminish too much over time. If you're like most people, however, your cooling is at its peak right now - or rather a few days ago when you installed the H50 - and it will get worse every hour you run your machine. Dust on the leading edges of fan blades, dust in the rad fins, etc., all will contribute to a decrease in cooling capacity.

If you're still using your computer two years from now, and you're not avid about cleaning the insides, then P95 may be simulating the conditions you'll have then when only stressing the CPU that 75%.

Click to expand...

Oh I'm pretty anal about that. Keeping things clean is keeping things cool and keeping things cool means more speed. I use dust off almost daily and once a month or so I shut down for a more thorough cleaning. I'll be keeping things clean for sure. I'm backing of to 3.7 though. I had to go right past 1.42 to 1.45v to get 3.8. Not worth it. 3.7 is stable at 1.4v An acceptable compromise.

I agree - I've lowered some of my rigs down from their stable top end just to keep things simple and cool.


As for the cleaning, good luck with the rad. I suggest a Q-tip and a LOT of patience.
My latest rad set-up will include filters - like for a HEPA air purifier. No good way to clean rads except removal followed by hot, soapy water (to the outside).

QuietIce said:

I agree - I've lowered some of my rigs down from their stable top end just to keep things simple and cool.


As for the cleaning, good luck with the rad. I suggest a Q-tip and a LOT of patience.
My latest rad set-up will include filters - like for a HEPA air purifier. No good way to clean rads except removal followed by hot, soapy water (to the outside).

Click to expand...

It won't take much to pull the thing off. I won't go pulling the water block but the fans and the rad are easy even to pull off. Then I can more effectively clean them all. When I do it I'll think of something I'm for cleaning the rad. There's just four long screws holding the rad on. R&R won't take but a few minutes and it's more than worth it. I am disabled/retired veteran though so I have more time than most.

Rigit said:

There's no issue though so I'll leave it alone. I don't like Prime95.

Click to expand...

That's the mentality I had then I would get a BSOD and reboot when gaming.