CPU overclocking problems. please help.

[Fefty_]

so i have a ryzen 7 2700 paired with a b450 aorus pro wifi board. even though i have a stock cooler, i decided just to see what juice i could get out of it. i ended off at 3.8ghz at 1.35v and it is stable through cinebench r20 tests. temps remain a few degrees off 70c. then, i decided to pump up the voltage to 4ghz with 1.4 volts, and even 1.42 and it barely made it through window boot, and black screened right away when i try cinebench. please explain to me why im crashing even on such high voltage (it's not like my cpu went extremely hot)??

[Jurrunio]

need 1.35v for even 3.8GHz? I know this board is garbage for overclocking but still.. that's a really bad result for a 2nd gen CPU, more like 1st gen behaviour.

 

btw if this is true, 4GHz is totally not going to work at 1.4 or 1.42V, the voltage jump from 3.8 to 4GHz is much more significant than that

[Fefty_]

8 minutes ago, Jurrunio said:

need 1.35v for even 3.8GHz? I know this board is garbage for overclocking but still.. that's a really bad result for a 2nd gen CPU, more like 1st gen behaviour.

 

btw if this is true, 4GHz is totally not going to work at 1.4 or 1.42V, the voltage jump from 3.8 to 4GHz is much more significant than that

forgot to check if 3.8 works for lower voltages but people have said they got the same chip up to 4.2 on 1.42volts. and wdym by "board is garbage", vrm temps seem very cool at under 50c monitored by HWiNFO64. is it a faulty reading? cuz people say this board goes over 100c when cpu OCed. also my psu is a 500w 80+ from antec, do you think the psu is bottlenecking the OC?

[Jurrunio]

4 minutes ago, Fefty_ said:

people have said they got the same chip up to 4.2 on 1.42volts.

your CPU isnt their CPU, silicon lottery is a thing an on Ryzen, could go as far as 300MHz between the best and the worst samples

 

5 minutes ago, Fefty_ said:

nd wdym by "board is garbage", vrm temps seem very cool at under 50c monitored by HWiNFO64.

1. Voltage regulation, i.e. how flat the voltage is, has nothing to do with VRM temperatures and can only be measured with a oscilloscope. The minimum voltage you need is actually the lowest spike in voltage, less spike = can run lower average voltage while staying stable, so better boards do better here. I was talking about this in my previous post

 

2. Cheap boards like this do not have VRM thermal sensor where you expect it to be so the reading should not be trusted, though if it crashes this quickly it's not the cause of the problem

[Fefty_]

so essentially i could say that the silicon lottery scammed me and even with a better cooler, my chip must cannot handle it? 

[SlayerOfHellWyrm]

3 hours ago, Fefty_ said:

so essentially i could say that the silicon lottery scammed me and even with a better cooler, my chip must cannot handle it? 

Not scammed, it's physics. Just because one  person, or even 90% of people can hit one OC with certain settings, doesn't mean you can. Silicon quality is going to vary from one part of the wafer your chip was made from, to another. To become a certain CPU, it only needs to be able to hit the advertised speeds and nothing more.

 

OC'ing, is running out of spec. As such, the silicon quality plays a heavy factor, and if you got the short end of the stick, your OC will either need higher voltage for the same clocks as someone else, or you just can't hit it. If you hit the lottery, you can do the same clocks with lower. As an example, my 8700K, I can get to 5.0 on all cores, off 1.35V, and 5.1 off 1.39 which is considered a bit lucky, since I think on average it needs like 1.4, or 1.41 to hit that.

Sometimes you win, sometimes you lose. This is why it's the silicon lottery.

 

[Fefty_]

thanks for your info..btw i decided to oc with ryzen master instead and surprizingly i could get 3.8 with 1.275v. i think there may be something wrong with my bios?