Any tips to quicker determine what hardware overclock is failing?

steffeeh

Reputable
Feb 12, 2016
265
1
4,815
You see my build in my sig

I'm currently experimenting with RAM settings and cache overclocking for the enjoyment (already having a CPU overclock that has been stable for a long, long time).
And lately my PC has been quite unstable.

At first I only thought that it probably was bad RAM settings, given that I pushed it to 2400MHz 10-12-12-28 1T 1.43V (with needed System Agent, VCCIO, and VTTDDR voltages), which is a quite hard push with those decreased timings.
But now it crashed at 2400MHz 13-16-16-28-2T 1.42V, which doesn't feel very logical, so I'm getting skeptic to if it's the RAM or not, as I'm getting close to the XMP settings (which is 2400MHz 14-16-16-31 2T 1.2V), yet it's still failing even when I still have a significant voltage increase - sometimes not too long into a session of heavier usage.

So I'm starting to think that maybe it's the CPU cache? I've upped it a bunch to 4.2GHz while keeping it at stock voltage, as I in the past had a long term stable CPU core overclock at 4.2GHz with stock voltage, which made me believe that I should expect the same from the cache.
However at the most recent crash with 13-16-16-28 etc, I had already bumped it down to 4.1GHz.
Also, the error messages I get on the mobo generally point towards the RAM, but I don't know...

I know I probably could just set it back to default and re-do it either RAM or cache first, wait for it to be stable a week or so, and then move on to the other one (as I probably should have done from the start), but perhaps there is a way to determine what overclock on what hardware is failing?
I've heard people suggest you can see of it's the CPU or RAM failing if the BSOD occurs instantly on a crash, or if the computer freezes first?

TLDR: Can I see what hardware OC fail depending on how the crash behaves? (strongly suggest you read the post though)

EDIT - MY BUILD:
Asus X99-A w/ BIOS 2101 | Intel i7 5820k OC @4.4GHz 1.264V w/ Noctua NH-U14S |
2x Corsair Vengeance DDR4 8GB 2400MHz 10-12-12-28 1T 1.4V | Asus Geforce GTX970
STRIX OC | EVGA 750 G2 750W | Samsung 850 Evo 1 TB | Windows 10 64-bit |
Be-Quiet Silent Base 800 w/ Silent Wings | 2x Dell U2414H OC @72Hz w/ Display
Port
(my signature doesn't display?)
 
Well, nobody told anything smart yet so i would just put ram to stock and test. If you still hav same issue disable your clock on cache.
Usually when cache unstable your screen will freeze and if RAM unstable it most often gives you bsod, and sometimes some warning in event viewer. Windows RMB>event viewer>windows log>system. At least this how it was for me.
 

steffeeh

Reputable
Feb 12, 2016
265
1
4,815
Usually, but not always? Meaning it's pretty much all about isolating what could cause the crashes I guess...
Any ideas on what voltages to use on the cache, is it similar to the core voltages?
 
You didnt explain how exactly your system is unstable. Do you get bsod? freeze? lag? During stress test or normal use? Gaming? What caused u to think that your system in unstable.
I cant say exactly what is going on from other side of the planet but at least i can try to help, this why i want to know info that i asked.
Make sure u made a thread on other forums too, u can try overclockers.com
 

steffeeh

Reputable
Feb 12, 2016
265
1
4,815
My bad, here you go :)

It mainly freezes for 3-4 seconds, and then I get a BSOD with the CLOCK_WATCHDOG_TIMEOUT error message. Sometimes I can't get out of the BSOD and have to press reset on the mobo.
This only happens in gaming, with games that uses realtime rendering rather than rendering when loading the saved world from the menu (as it does in many FPS games for instance).
I will play for example Ark or Just Cause 3, for 1.5 hours or so, before a crash MAY happen (not always), or in Minecraft immediately when it renders the world the first time when loading a saved game - if it doesn't crash there, it pretty much won't crash during a session.
This is despite that I've run a complete system stability test in Aida64 for 2+ hours.

For now I've set the RAM to XMP settings and upped the cache back up again, to see if it crashes again (that would indicate the cache is unstable).
For the moment I'll only keep the thread here and see where it gets me, might post on overclockers.com and elsewhere if needed (I try to post there as little as possible since they can be a bit harsh there).