That's not at all always true. Stock and default settings of some vendors auto clock out of the box...and are far from optimal in some cases and by so wiill degrade your cpu worse than doing an proper undervolt. From personal experience ASUS motherboards on the 5th gen Intel if IIRC...
Well sorry to hear you have had terrible luck with all the CPU's you had. I have had some poor silicon in the years but mostly it have worked out fine.2DaT wrote: Wed Jul 06, 2022 3:49 amWell, I tried to do some overclocking in 2010s and had terrible luck with all CPUs I had... Silicon lottery is a thing, and you may not win it. These days CPUs have even less headroom than they did before. And people on the internet overestimate the clocks, or stability of the system very often.Kr3eM wrote: Wed Jul 06, 2022 1:54 am It's not a question about outsmarting the department. They could probably lower the wattage by 10% or perhaps even more, but there would be less die's on each wafer that would be of that quality....
There is always a risk of unstable operation, that's what you need to acknowledge when doing overclocking/undervolting. It may be not obvious, it could be rare crash once a month, that doesn't reveal itself on stress test. That's #1 reason why I advise against it.
My current rig is a decent OC i2500k and have been running stable as a rock for 8+ years...... and still no signs of degradation. The only major unstable issue I ever got was due to a buggy update of a plugin, the system was heavily tested when I overclocked it under a extrem heatwave so it was during the worst conditions, after that I was pretty certain and it was quite easy to track down any error when only doing manual updates...it would have less work if I had not update so many plugins at once thoug...
Pretty much the same with my daily gaming rig Ryzen 2600 but I only choose to run the my custom OC profile when I need a little extra in some games...when it makes an actuall difference, sure there's less headroom these days, but it's not always about having higher clocks. It seems that it got more common with all the default boost and auto functions in BIOS to run things with much higher voltage than they actually might need and sometimes it can be a good idea to see if one's setup really need that much voltage for the that amount of performance... I mean if one are concerned with degradation over time.
My Linux 5600G daily setup is slightly undervolted and pulls merly 30w in most conditions be it streaming 1080p video or 480p popup while running 2 browser with loads o tabs and what not and idles about 18w....thats for the whole system pull from the wall.
If there's a random crash once a month then i would rather question what auto updates and background tasks and services is running on the system, specially when working on a project...if it's something that is unstable due to OC then it should be easy to replicate, well it should also be less of a problem if it's due to a plugin or software but the more things running by itself doing whatever they like whenever...well then tracking down the cause gets much more problematic...every update is a potential risk of a bug being introduced.
Personally I would never advice anyone not doing some research about their gear and what can possibly be done, if it peaks their interesse.... if one don't succed one can just go back to stock and in the end one can have at least learned something about how the BIOS works...

