[SOLVED] My first RYZEN overclock

yui_nitsu

Prominent
Oct 15, 2017
8
0
510
I have a ryzen 5 1400 and i tried to OC it to 3,5, setting the voltage to 1.18~1.20v, i dont know if this voltage is good for the CPU. The temperature is fine but i still need to make some stress tests.

Here is some pics of the cpuz and howmonitor.
https://i.imgur.com/RLgD8SE.jpg

Is everything fine?
 
Solution
What motherboard and cooler do you have?

Everything is just fine! The voltage is extremely low... but then that is a very mild overclock too. Don't be afraid to increase voltage. The processor will be perfectly fine at 1.4 volts; AMD recommends 1.425 as the upper limit long term. At higher voltage you could start pushing clock speeds up towards 3.8Ghz, 3.9G; maybe even 4.0G.

Overclocking Ryzen is really very easy, it's mostly a matter of balancing voltage and clock speed to keep temperature (while stress testing) below 80C or so. Just so long as you don't set a voltage above 1.4 (1.425 max) you'll be OK. So it depends heavily on your cooler; Ryzen loves good cooling to hit higher clocks.
What motherboard and cooler do you have?

Everything is just fine! The voltage is extremely low... but then that is a very mild overclock too. Don't be afraid to increase voltage. The processor will be perfectly fine at 1.4 volts; AMD recommends 1.425 as the upper limit long term. At higher voltage you could start pushing clock speeds up towards 3.8Ghz, 3.9G; maybe even 4.0G.

Overclocking Ryzen is really very easy, it's mostly a matter of balancing voltage and clock speed to keep temperature (while stress testing) below 80C or so. Just so long as you don't set a voltage above 1.4 (1.425 max) you'll be OK. So it depends heavily on your cooler; Ryzen loves good cooling to hit higher clocks.
 
Solution


My motherboard is a B350ET2 and my cooler is a Hyper T2.

Should i worry about the VID numbers? 1.550 V seems strange, i think its a bug
 


VID is the part of the CPU's internal control scheme that's been taken out of the loop when you put the processor into manual overclocking mode by setting a fixed multiplier and voltage. It's not relevant when overclocking, and is often pegged at 1.55V in many boards when you do.

The voltage to watch is "CPU VDD (Node 0)" in HWMonitor. That's the actual core voltage reported by the processor in telemetry. CPU Vcore is (usually) reported by the VRM controller and is often quite a bit higher than CPU VDD, especially when under load, due to VDroop. In better boards it will be measured close to the CPU so VDroop is considered and will be a lot closer to CPU VDD (Node 0).

As a strategy for safety: if you keep CPU Vcore below 1.425 Volts the processor will never see a harmful voltage range since VDroop always drops it even further. For even further margin, use 1.4V as your max CPU Vcore. And don't be alarmed if you should set voltage a bit high and and see high excursions, even above 1.45 volts. It's safe for short bursts and people regularly run at 1.5V for a quite some time while benching. But for longest assured life, just reduce voltage so it doesn't go above 1.4-1.425V and proceed!

Hyper T2, while pretty good, is still a limited cooling solution. Not to sure how far you can go, even though it's only a 4 core CPU, before it gets saturated and temperature starts to climb.
 


Should i worry about the Currents numbers? When i'm gaming, it goes to 32.22 A
https://i.imgur.com/udR1erU.jpg
 


I don't see how you can consider them meaningful as absolute values, but even so 32A is a minor current load on a VRM. Remember, you've got a 4 core...an 8 core Ryzen 1st gen at 4.0Ghz will supposedly draw about 100A in heavy loads. But my 8 core can draw over 120A, as reported, and it's only operating at 3.9Ghz. I'm not sure the system is properly calibrated to report current or watts accurately.

The best way to determine power draw is to get a watt meter (like a Kill-A-Watt) and plug the computer into it and then into the wall. By measuring and comparing differential values, idle vs full load, un-overclocked vs. overclocked, you can better assess power consumption.