overclock crashed after 1hour stress test. can i still use it though?

henkkrakers

Commendable
Mar 5, 2017
66
0
1,640
so i overclocked my cpu but it ran in aida 64 for over an hour before it crashed. since my pc isn't going to be used at 100% cpu usage for anywhere near this long is it safe to say that this is a good enough overclock to stay stable under normal load. (i won't be rendering or anything like that, only gaming)

if i play games and my cpu usage is say 50% what is the likelihood of it crashing during this? (if it's like 0.0001% chance of happening im fine with it. if it's 1 crash every hour obviously thats not ok.

also I have gone to 1.45v which is amd's max recommended voltage on the vcore. what happens if i go higher? the temps stay under 75c which is well below thermal max and that's at 1.45v.

nb soc voltage is 1.1v at the moment. thinking of putting it too 1.25v like someone else did.

i know downclocking 100mhz is the smart choice however i would prefer to max out my chip unless it's absolutely not feasible. any tips/advice, thanks

cpu=1700x
mobo= MSI X370 GAMING PRO CARBON
ram = cheap 2400mhz 2x8gb ram from crucial. (still not cheap lol)
 
Solution
Both would contribute. Frequency would be what affects the probability of an error for each operation, and the higher the CPU load the more operations the CPU is performing per second, and therefore higher load means higher chance of error per unit time.

So running at max frequency and 50% load should result in less errors on average than 100% load. The type of load can affect it too, which is why you can sometimes be stable with one stress test for an hour and then crash in 10 minutes with another.

henkkrakers

Commendable
Mar 5, 2017
66
0
1,640


is it about hitting the core frequency that triggers the crash or is it the cpu load (e.g. 100% load) that causes a crash. which one is more likely to cause a crash? in games i imagine the frequency will always be as high as possible whereas load is usually not ever 100%

 

TJ Hooker

Titan
Ambassador
Both would contribute. Frequency would be what affects the probability of an error for each operation, and the higher the CPU load the more operations the CPU is performing per second, and therefore higher load means higher chance of error per unit time.

So running at max frequency and 50% load should result in less errors on average than 100% load. The type of load can affect it too, which is why you can sometimes be stable with one stress test for an hour and then crash in 10 minutes with another.
 
Solution

henkkrakers

Commendable
Mar 5, 2017
66
0
1,640


ok thanks. i guess the only thing to do is test it for a few weeks and see what happens