Can someone explain this CPU overclock situation

[lafrente]

I all core locked my R5 3600 to 3.8ghz since I didn't like it jumping up and down all the time.

And I locked the volts to 1.150. I know pretty low volts, I don't need the performance.

 

So, static 3.8ghz and 1.150v 

 

Here is cinebench result:

 

Then I kept the voltage at 1.150v, and raised the clock to 4ghz.

Here is the better result

 

What's going on here? I feel like going 4.1 and testing it again.

[IAmNik]

If youre stress testing, dont you want the extra volts? yu said yu dont need the performance but doesnt a stress test test the performance? please, correct me if im wrong here and seeing the pics wrong. I dont benchmark alot lol

[lafrente]

8 minutes ago, IAmNik said:

If youre stress testing, dont you want the extra volts? yu said yu dont need the performance but doesnt a stress test test the performance? please, correct me if im wrong here and seeing the pics wrong. I dont benchmark alot lol

I want low volts since I dont need the performance and I dont want the fans ramping due to heat, but I also would like to know what is the max possible speed at this voltage.

[IAmNik]

6 minutes ago, lafrente said:

I want low volts since I dont need the performance and I dont want the fans ramping due to heat, but I also would like to know what is the max possible speed at this voltage.

Ahhh ok, i get what youre saying. Like I mentioned, I dont really do any benchmarks, just figured that extra voltage could help during the benchmark itself.

[Biohazard777]

23 minutes ago, lafrente said:

What's going on here?

You increased the clock and hence got a better result, no surprise there.
As for why your CPU is stable at 1.150V and 4.0 GHz... 1st test if it really is stable as undervolting can fall apart at both low and high loads, see if your system is truly stable at 4Ghz and that voltage (use the PC for a few days as you normally would).

 

10 minutes ago, lafrente said:

but I also would like to know what is the max possible speed at this voltage.

That you will have to test out yourself. Because no two CPUs are exactly the same, you might have gotten lucky at the "silicon lottery".

 

[lafrente]

1 minute ago, Biohazard777 said:

You increased the clock and hence got a better result, no surprise there.
As for why your CPU is stable at 1.150V and 4.0 GHz... 1st test if it really is stable as undervolting can fall apart at both low and high loads, see if your system is truly stable at 4Ghz and that voltage (use the PC for a few days as you normally would).

 

That you will have to test out yourself. Because no two CPUs are exactly the same, you might have gotten lucky at the "silicon lottery".

 

Wow that's weird, I used this CPU for a long time 3.8 static at 1.150v just thinking that was low enough already. I tested for a few hours, nothing is breaking down for now.

 

Do you think I harmed it in any way by giving it enough voltage to work at 4ghz and making it work at 3.8? 

[Biohazard777]

2 minutes ago, lafrente said:

Do you think I harmed it in any way by giving it enough voltage to work at 4ghz and making it work at 3.8? 

You weren't using its full potential, but no I don't think you were harming it. Only if you are feeding it seriously high voltage (like 1.4V or more) then I'd be concerned about the lifespan (it can be shortened significantly), you were doing 1.15V and that isn't high for Ryzen 5 3600. As you already know less voltage = less wattage (for the same amount of current drawn) = less heat = longer lifespan. Now yeah maybe you could have hit 3.8 Ghz with even less voltage, like 1.05V but IMO it wouldn't matter too much. How much more lifespan you'd be adding to the components could be negligible... like will it run for 14 or 16 years kind of question heh (pulled those two numbers out of thin air, we are yet to see how long Ryzen CPUs will last, though generally speaking CPUs are one of the longer lasting components). Same goes for the mobo and the PSU which supply and regulate the power to the CPU.... maybe you could have squeezed out a bit more time out of them, but in the end it doesn't really matter, the whole system will long be obsolete by then.

How high / low can you go... depends on how lucky (or unlucky) you got. For example. some can hit all core 4.2 GHz @ 1.1V and some need 1.35V.