Healthy overclock?

[oakaxe]

I just recently built my first computer thanks to memory express's price beat guarantee (thanks ME) and I was overclocking my Ryzen 3 2200G (cpu and gpu) and whenever I stress tested it (cpu) at 2900Mhz it would reach about 105C. I did not think that this was a quite healthy so I lowered it to 2800Mhz but whenever I am in games it never goes any higher than 50C. Maybe it's because I am not playing any highly demanding games (borderlands 2, PC building simulator, Fortnite, BTD6) but should I raise it to 2900Mhz anyways? The stress test said that there were no errors or problems I simply saw a 105C temp and thought it best to go back a bit.

 

Update: Using RealBench, I was able to test my cpu and figure out a stable 4Ghz oc on all cores. Thanks to everyone who helped me out. 

Update: RESOLVED

Edited May 7, 2019 by oakaxe
issue resolved

[Fasauceome]

Depends a lot in what you're using to monitor the temps

[Master Disaster]

What stress test did you run?

[GOTSpectrum]

If you wanna check on long term temps try getting involved in our BOINC event and give back to science while you are at it. 

 

 

[oakaxe]

6 hours ago, fasauceome said:

Depends a lot in what you're using to monitor the temps

 

6 hours ago, Master Disaster said:

What stress test did you run?

I was using Ryzen master to monitor temps while running prime 95. I was running a Small FFT's test for about 5-10 mins at a time with a jump of 100 mhz between every test.

[oakaxe]

1 hour ago, Ben Quigley said:

If you wanna check on long term temps try getting involved in our BOINC event and give back to science while you are at it. 

 

 

How would a BIONIC event help me test my cpu?

[GOTSpectrum]

6 minutes ago, oakaxe said:

How would a BIONIC event help me test my cpu?

It puts it under 100% load for long periods of time... if your OC isnt 100% stable you'll fail WUs

[Master Disaster]

11 hours ago, oakaxe said:

 

I was using Ryzen master to monitor temps while running prime 95. I was running a Small FFT's test for about 5-10 mins at a time with a jump of 100 mhz between every test.

Small FFTs is a very unrealistic test, it's actually a torture test more than a stress test. It's purpose is to push as hard as possible so the temps you get while running it are not indicative of anything really. It's designed to test stability, not temperature.

 

Use blend for a much more accurate to real world result. The maximum temperature you see in a blend test will be around the maximum you see in 95% of day to day scenarios.

[oakaxe]

5 hours ago, Master Disaster said:

Small FFTs is a very unrealistic test, it's actually a torture test more than a stress test. It's purpose is to push as hard as possible so the temps you get while running it are not indicative of anything really. It's designed to test stability, not temperature.

 

Use blend for a much more accurate to real world result. The maximum temperature you see in a blend test will be around the maximum you see in 95% of day to day scenarios.

Oh ok thank you :) 

[steelo]

2 hours ago, oakaxe said:

Oh ok thank you

I really don't like Prime95 personally, I can't see many real world scenarios where everything is absolutely pegged to the max and it just severely tortures hardware. Realbench is supposed to be a good stability/stress test, if you can get that to run for at least 15 minutes without reporting instabilities, you are usually good to go.

[oakaxe]

7 hours ago, steelo said:

I really don't like Prime95 personally, I can't see many real world scenarios where everything is absolutely pegged to the max and it just severely tortures hardware. Realbench is supposed to be a good stability/stress test, if you can get that to run for at least 15 minutes without reporting instabilities, you are usually good to go.

I'll test that out then. I only used Prime 95 since it was recomended on the oc article that I was reading.

[steelo]

On 5/6/2019 at 7:33 PM, oakaxe said:

I'll test that out then. I only used Prime 95 since it was recomended on the oc article that I was reading.

I'm sure there are a lot of experts out there who like it. I, personally don't like P95 because it essentially puts hardware through a lot of stress that 99% of the time it never would encounter. Also, I don't like running my CPU at 80-90c for any duration.

 

Heaven is another good one for stability testing your GPU and CPU. Just run it on a loop for an hour or so.