i7-4770K strange unstability at 4.1 or 4.2

[atarione]

So i have a Gigabyte Z87X-UD3H /i7-4770K /16GB corsair DD3

I have been messing about trying to overclock this thing and something weird keeps happening... 4.1 or 4.2Ghz(1.25vcore) pass bench marking / stress tests (8hrs XTU stress test) and 4~hrs or so with RealBench

however... when overclocked (and only when overclocked) about 1/2 the time the windows start menu and search fail to work and when shutting down or logging out get a message .. despite reference to reboot required signing out and back in seems to resolve problem (frequently ..not always)..



looking at the event logs this seems "suspect" I don't know if this stupid little Bluetooth adapter is choking when the system is overclocked.. kinda ran out of time to mess around with this again atm... I guess I'll try taking the Bluetooth adapter out later and seeing if the problem goes away?

The CSR Bluetooth Audio Service service did not shut down properly after receiving a preshutdown control.

Any Tips / ideas???

 

[atarione]

So 1/2 dozen restarts later after I got home with the bluetooth adapter removed the problem isn't happening.. so I guess why doesn't the bluetooth adapter liking the oc?? only changed the cpu multiplier and added .005 offset voltage??

 

[atarione]

nope never mind problem came back the bluetooth adapter removed.. no windows start button , no search and the plug and play message on logoff or on.

 

[Pandur]

Try running memtest to make sure your memory controllers are not misbehaving?

Check your secondary voltages and cache multiplier. Maybe your chipset voltage needs a tiny boost or similar. Maybe start by listing what your cache multiplier is set at and what your other voltages are at (other than vcore).

 

[atarione]

Hi... so I don't actually think it was the OC.. problem showed up at stock speeds .. after much head to desk frustrations it seems that issue tracks back to a "creators edition" non sense..

finally found "work around" .. turn off "Use my sign-in info ti automatically finish setting up my device after an update or restart" in sign on options.. apparently this feature causes the issue I saw before for some users.. ...

So far all i had done was change the multiplier to x41 or x42 and apply .005+ offset voltage to the Vcore.. which may in fact be stable based on my earlier results (excepting the weird sign off message). Seems Microsoft didn't do a bang up job testing the creators update?? meh...

 

[prime2515102]

Hi... so I don't actually think it was the OC.. problem showed up at stock speeds .. after much head to desk frustrations it seems that issue tracks back to a "creators edition" non sense..

finally found "work around" .. turn off "Use my sign-in info ti automatically finish setting up my device after an update or restart" in sign on options.. apparently this feature causes the issue I saw before for some users.. ...

So far all i had done was change the multiplier to x41 or x42 and apply .005+ offset voltage to the Vcore.. which may in fact be stable based on my earlier results (excepting the weird sign off message). Seems Microsoft didn't do a bang up job testing the creators update?? meh...

That is really bizarre. It doesn't surprise me though, I've had all kinds of weird things happen after the creator's update (like randomly not being able to switch weapons or use the fire button in Doom... wtf?).

 

[spine]

Same platform as mine, and IIRC, Gigabyte boards just need extra voltage to the CPU Eventual Input voltage to equal Asus boards with this chip in terms of overclockability.

I've had this board for over a year now, ~5ghz OC, and for Asus, 2.1v gets you there, but reportedly, the same chip on a 'given' Gigatbyte, and you're looking at more like 2.20-2.30v.

There's an extensive thread here...
http://www.overclock.net/t/1411077/haswell-overclocking-guide-with-statistics

I've followed that thread since inception, so any Q's RE Haswell OC....

 

[atarione]

thanx for the link... so in my case some "confusion" occured due to some bugs the recent fall creators update to window 10 was causing ..that I first assumed had been overclock related.

so far /touch wood the 4.2Ghz overclock is stable and temps seem ok 68~ on hottest core running XTU stress test or realbench for awhile (not long enough to be sure yet..) but XTU ran 1/2 hur with no troubles.

I will read through the links provided.. I can't decide if I wanna go higher right now or not?

I'm also debating getting a new cooler ... I do not want to get an AIO because I'm too worried about leaks or pump failures (i do have a few friends that have had one or the other happen)

I'm considering the cyorig H5 as it is reasonably cheap @$42~ish dollars on amazon and because it isn't so gigantic as to be a complete pain in the arse while still sounding like a decent improvement over the 5~yrs old Hyper212+ I have now?

thoughts?

oh here is the realbench benchmark results 4.2Ghz w HD7950 (950/1400)