Question Can anyone explain this please

Fiorezy

Great
Jul 3, 2020
116
14
95
2
When I stress my CPU all cores goes up to 4000Mhz and temps to around 78c, but when gaming it goes to around 4250Mhz and the temps are 62-65c. Can you explain please?

My setup:
MSI MPG x570 Gaming Plus
Ryzen 7 3700x
RTX 2060 Super
16GB DDR4 2666Mhz
SSD SATA

Nothing is overclocked, XMP, stock cooler
 

kanewolf

Titan
Moderator
When I stress my CPU all cores goes up to 4000Mhz and temps to around 78c, but when gaming it goes to around 4250Mhz and the temps are 62-65c. Can you explain please?

My setup:
MSI MPG x570 Gaming Plus
Ryzen 7 3700x
RTX 2060 Super
16GB DDR4 2666Mhz
SSD SATA

Nothing is overclocked, XMP, stock cooler
Yes. Single core boost is higher than all core max clock. So when gaming, you have less than all 8 physical cores maxed. Therefore 1 or 2 cores can go to a higher clock speed.
 

Fiorezy

Great
Jul 3, 2020
116
14
95
2
When you stress test all cores to max, temperature lowers frequency, 78c is just too much for it. Ryzen is best at up to 62-65c and that's why you see higher frequency under game load,
Ryzen 2000 and 3000 series thermal trottle at 85c, so 78c is is still fine
 
Could you please provide a source for this?
Source is my system and extensive testing with several coolers. First heard about it here (I think)
View: https://www.youtube.com/watch?v=g94rNe4XSGU
didn't exactly believed it at first so I experimented too and results were roughly same not counting silicone lottery.
My results with 3 consecutive runs(enough to stabilize temps) of CB r20 .
62c - 5024 (highest boost 4.382MHz one core at a time and 4.25GHz other cores
71c - 4963, 4.31Ghz best core, 4.2 other cores.
75c - 4892, 4.28GHz best core, 4.1other cores
78c -4797. 4.125GHz best core, 3.97 other cores
There are some other scores in between and at a bit higher temps but there's too many to list here.
I also had 1600x, 1700x (with XFR only) and 2700x previously. They were on x370 MB and 2933 and 3000MHz RAM so can't compare them with present system and I also lost results due to a virus at one time.
Who doesn't believe let him try himself.
PS. I never got them to throttling point, those are just frequency changes impacting PBO and test results with it. Temps checked with HWinfo64 and Ryzen Master.
 

Fiorezy

Great
Jul 3, 2020
116
14
95
2
Source is my system and extensive testing with several coolers. First heard about it here (I think)
View: https://www.youtube.com/watch?v=g94rNe4XSGU
didn't exactly believed it at first so I experimented too and results were roughly same not counting silicone lottery.
My results with 3 consecutive runs(enough to stabilize temps) of CB r20 .
62c - 5024 (highest boost 4.382MHz one core at a time and 4.25GHz other cores
71c - 4963, 4.31Ghz best core, 4.2 other cores.
75c - 4892, 4.28GHz best core, 4.1other cores
78c -4797. 4.125GHz best core, 3.97 other cores
There are some other scores in between and at a bit higher temps but there's too many to list here.
I also had 1600x, 1700x (with XFR only) and 2700x previously. They were on x370 MB and 2933 and 3000MHz RAM so can't compare them with present system and I also lost results due to a virus at one time.
Who doesn't believe let him try himself.
PS. I never got them to throttling point, those are just frequency changes impacting PBO and test results with it. Temps checked with HWinfo64 and Ryzen Master.
That is interesting! What was your room temp when you got 62c?
 

Fiorezy

Great
Jul 3, 2020
116
14
95
2
Not really, it's built into CPU. There are some settings in my BIOS for TDC and shutoff temps but didn't see any reaction to them.
Thanks you for your replies,
I already tweaked PBO and voltage and got some improvement, Cinebench R50 is now scoring 5000pts with 75c, gaming which is my main task is now 60-63c @4200-4300MHz. I'm very satisfied with this
 

ASK THE COMMUNITY