F5F Stay Refreshed Power Users Overclocking RYZEN 3600 OC - advice

RYZEN 3600 OC - advice

RYZEN 3600 OC - advice

Pages (3): Previous 1 2 3
N
NotLulu
Member
66
12-29-2019, 09:18 AM
#21
True. As I previously discussed, I will maintain these CPU settings: 1.24 V with a 42 GHz clock speed (temperatures in Prime reach a maximum of 72.5 degrees, which seems acceptable). Also, I won't even encounter this kind of workload in any game.
N
NotLulu
12-29-2019, 09:18 AM #21

True. As I previously discussed, I will maintain these CPU settings: 1.24 V with a 42 GHz clock speed (temperatures in Prime reach a maximum of 72.5 degrees, which seems acceptable). Also, I won't even encounter this kind of workload in any game.

C
coolman4566
Member
61
01-04-2020, 11:39 PM
#22
One question, how long have you been adjusting the settings you claim to use?
Have you tested Realbench, prime95, aida64, etc., for extended periods to confirm stability?
Just because a CPU can handle Cinebench and prime95 briefly without problems doesn’t guarantee reliability.
If you want a truly stable CPU quickly, I’d perform a small FFT test with AVX disabled (and monitor closely the number of workers—stop when one worker drops, indicating insufficient voltage or too high frequency). The blend test should be saved for later when you’re confident your CPU is stable.
If your Ryzen 5 3600 runs rock-solid at 4.2GHz @ 1.24v, you’re extremely lucky—truly a rare gem.
You essentially receive the maximum advertised turbo boost speed from the poorest Ryzen 3rd generation CPU (excluding the 3500X), which is just incredible (and unbelievable for most people with a R5 3600, including me).
C
coolman4566
01-04-2020, 11:39 PM #22

One question, how long have you been adjusting the settings you claim to use?
Have you tested Realbench, prime95, aida64, etc., for extended periods to confirm stability?
Just because a CPU can handle Cinebench and prime95 briefly without problems doesn’t guarantee reliability.
If you want a truly stable CPU quickly, I’d perform a small FFT test with AVX disabled (and monitor closely the number of workers—stop when one worker drops, indicating insufficient voltage or too high frequency). The blend test should be saved for later when you’re confident your CPU is stable.
If your Ryzen 5 3600 runs rock-solid at 4.2GHz @ 1.24v, you’re extremely lucky—truly a rare gem.
You essentially receive the maximum advertised turbo boost speed from the poorest Ryzen 3rd generation CPU (excluding the 3500X), which is just incredible (and unbelievable for most people with a R5 3600, including me).

C
Coolncreamy
Junior Member
25
01-08-2020, 10:29 PM
#23
executed run prime95 with small FFt, turned off the three boxes for AVX-512 and AVX2. this configuration will fully utilize your CPU. my 8700k (5GHz @ 1.255v) reaches a maximum of 78c during this test, while gaming it stays between 50 and 60 dephending on the game.
C
Coolncreamy
01-08-2020, 10:29 PM #23

executed run prime95 with small FFt, turned off the three boxes for AVX-512 and AVX2. this configuration will fully utilize your CPU. my 8700k (5GHz @ 1.255v) reaches a maximum of 78c during this test, while gaming it stays between 50 and 60 dephending on the game.

E
Elina_Aada20
Member
141
01-13-2020, 08:18 PM
#24
Ok. I am one hour in the Prime test you have suggested. So far It is passing all the test on all workers althouth there is one thing which is strange: On worker for example number 5 Self-test 16k passed at 10:46 but on worker number 3 it passed the same test at 10:44. I don't know whether that is normal or not. I am new to Prime95
I will post some screenshot as well later.
E
Elina_Aada20
01-13-2020, 08:18 PM #24

Ok. I am one hour in the Prime test you have suggested. So far It is passing all the test on all workers althouth there is one thing which is strange: On worker for example number 5 Self-test 16k passed at 10:46 but on worker number 3 it passed the same test at 10:44. I don't know whether that is normal or not. I am new to Prime95
I will post some screenshot as well later.

D
danieeel
Junior Member
6
01-14-2020, 01:14 AM
#25
The test has completed successfully after two hours. Everyone was working properly. The screenshot of HWinfo is attached.
What should I do next? Should I consider running the blend test?
Thank you for your responses!
D
danieeel
01-14-2020, 01:14 AM #25

The test has completed successfully after two hours. Everyone was working properly. The screenshot of HWinfo is attached.
What should I do next? Should I consider running the blend test?
Thank you for your responses!

S
Skulhead2a
Member
210
01-21-2020, 02:14 PM
#26
If that evaluation was based on the FFT and there are no AVX optimizations, you can proceed to other tests to verify your OC.
Any changes in voltage and frequency are straightforward, but the most time-consuming aspect is the testing duration needed to confirm.
I didn’t think my OC was stable until it could achieve P95 for 8 to 10 hours. This included the Realbench Stress test—a 8-hour session using half your RAM, OCct, Aida64 extreme stress tests with CPU, cache, and RAM boxes monitored for 8-10 hours—followed by actual gaming hours to check for issues.
Good luck!
S
Skulhead2a
01-21-2020, 02:14 PM #26

If that evaluation was based on the FFT and there are no AVX optimizations, you can proceed to other tests to verify your OC.
Any changes in voltage and frequency are straightforward, but the most time-consuming aspect is the testing duration needed to confirm.
I didn’t think my OC was stable until it could achieve P95 for 8 to 10 hours. This included the Realbench Stress test—a 8-hour session using half your RAM, OCct, Aida64 extreme stress tests with CPU, cache, and RAM boxes monitored for 8-10 hours—followed by actual gaming hours to check for issues.
Good luck!

Pages (3): Previous 1 2 3