Is it okay for Ryzen 3000 auto voltage to have such stark temp and voltage changes?

[GregI]

While I'm gaming, I'll notice my 3600 has such drastic temperature changes. One second it'll be 43c and then next refresh on screen it's 51c. Next second 48c then 43c and back to 50c.. you get the point

I've just looked at CPU voltage in CPU-Z and is it really ok to have stuff like this happen every other second:

https://i.imgur.com/Q9pS3KV.png
https://i.imgur.com/mGFBiz6.png

I wonder if it's better to go into the BIOS and go about setting a standard voltage or if CPUs are so advanced now temp and voltage changes like this don't matter at all? I am running the AMD Ryzen High Performance Power Plan if that makes a difference

 

[Mcnumpty2323]

Been told that voltage is normal if its under lower load
But high load it drops
But you can set 1.3v in ryzen master
Or bios if you want to
I tried 1.3v in ryzen master seemed to cause no problems

 

[Panos]

While I'm gaming, I'll notice my 3600 has such drastic temperature changes. One second it'll be 43c and then next refresh on screen it's 51c. Next second 48c then 43c and back to 50c.. you get the point

I've just looked at CPU voltage in CPU-Z and is it really ok to have stuff like this happen every other second:

https://i.imgur.com/Q9pS3KV.png
https://i.imgur.com/mGFBiz6.png

I wonder if it's better to go into the BIOS and go about setting a standard voltage or if CPUs are so advanced now temp and voltage changes like this don't matter at all? I am running the AMD Ryzen High Performance Power Plan if that makes a difference

Leave the voltage to default. Don't tamper with it. Thats the correct behaviour of the CPU.
What you need to do to get the extra out of your system, is use the 1usmus power plan. Use the 1.1 version but look at the 1.0 document for how to setup up your BIOS.

Do not try manual OC this CPU let alone try to push 1.3v+ manually. Under normal conditions can use up to 1.5v however that's within the AMD safe envelop. Anything over 1.27v for all core 100% load overclock is risky and over 1.3v expect the CPU to degrade within 6 months and die.

Again this is for manual overclocks, leaving everything default and using PBO or Auto OC is perfectly fine. The CPU knows what's safe and what not.

 

[Drumroll]

Normal.
The voltages, unless setting a static voltage, like when overclocking, fluctuate. No issues.
The temperature spikes, same thing. Unless performing a task for long, th cooler won't be able to avoid it. Not an issue. What happens is those spikes happens at fractions of seconds. The response from the motherboard controller, increasing the fan's rpm won't, ever, be as fast.
For a more steady scenario, setting a static voltage, would return more predictable temps and voltage, but won't allow some energy saving features.

 

[andy_mk3]

Perfectly normal, it's how it is designed to function. Don't force it to lower voltages you will lose performance.

 

[Deleted member 66701]

While I'm gaming, I'll notice my 3600 has such drastic temperature changes. One second it'll be 43c and then next refresh on screen it's 51c. Next second 48c then 43c and back to 50c.. you get the point

Normal.

I've just looked at CPU voltage in CPU-Z and is it really ok to have stuff like this happen every other second:

Yes, perfectly normal.

I wonder if it's better to go into the BIOS and go about setting a standard voltage

No, don't do that.

or if CPUs are so advanced now temp and voltage changes like this don't matter at all?

Nail meet head.

 

[GregI]

What you need to do to get the extra out of your system, is use the 1usmus power plan. Use the 1.1 version but look at the 1.0 document for how to setup up your BIOS.

How much performance am I likely to gain from doing this for gaming? It also seems to be something for the higher end Ryzen 9 processors though?

 

[Panos]

How much performance am I likely to gain from doing this for gaming? It also seems to be something for the higher end Ryzen 9 processors though?

use v1.1 universal, on anything over 3600 would work

 

[Joxeon]

Do not try manual OC this CPU let alone try to push 1.3v+ manually. Under normal conditions can use up to 1.5v however that's within the AMD safe envelop. Anything over 1.27v for all core 100% load overclock is risky and over 1.3v expect the CPU to degrade within 6 months and die.

Do you have any evidence of this or are you just making it up? I've been running my CPU at 4.2 all core for the last 6 months now at 1.325v and there is no deg, infact it runs 5c cooler than stock and 10c cooler than pbo at full load. Imo pbo is more dangerous as at full load it tries to feed the chip 1.37v hence the extra temps.

 

[Killer7xx]

This is normal behaviour, the third gen Ryzen CPUs have a very high resolution boosting algorithm which can ramp the speed up and down in a fraction of a second in response to a changing load on the CPU. For the majority, it's better to use stock voltage and settings since undervolting can starve the CPUs of power and cause performance drops or stability issues in some cases.

Adding more voltage can also cause the CPUs to use more power and get hotter. You can OC and add voltage if you want higher multicore perf, but keep in mind that you'll lose some single core performance since it will lock all cores at the same max freq and the stock single core boosts being largely unachievable across all cores.

If you want more performance you should look at turning on PBO and the auto OC settings. That'll ensure more aggressive clocks across all cores in multicore workloads whilst keeping a similar or better single core boost to stock. Personally I don't use it myself so I'm not that clued up on how to set it up, but there are guides out there that explain it in more detail.

 

[GregI]

Okay I might have messed this up somewhere. I was doing some setting changes and now I have noticed CPU-Z will not show me a voltage below 1 volt anymore. It's all between 1.2 and 1.4

What windows setting or BIOS setting is responsible for this?

 

[thesmokingman]

While I'm gaming, I'll notice my 3600 has such drastic temperature changes. One second it'll be 43c and then next refresh on screen it's 51c. Next second 48c then 43c and back to 50c.. you get the point

Are you using Afterburner, ie. poll based monitoring? If yes, then switch to clock based monitoring instead. Polling software creates the issue itself by constantly polling the cpu causing the spikes and dips. If you want to look at a cpu temp use the die averaged sensor, otherwise with AB/RTSS you see individual core temps instead of an averaged temp.

https://www.techpowerup.com/257312/...age-exaggerated-a-case-of-the-observer-effect