Overclock became unstable overnight! R4E & i7 4930k

[terente]

I finally decided to upgrade my aging system.
Bought a new Rampage 4 Extreme motherboard with EK watercooling kit.
i7 4930k CPU, retail, watercooled.
32Gb Corsair Quad channel RAM CMZ32GX3M4A1866C10 (1.5V kit)
EVGA Hydro Copper 780Ti - GPU
Zotac GTX480 watercooled - Physx

There's plenty of cooling, it's a three loops system with 5 480 rads and 7 laing ddc pumps.

First encounter with the BIOS, a month ago I loaded the automatic profile for 4.5GHz and DDR 1600, dialed CPU voltage back @ 1.35V (from automatic 1.4V) and DDR voltage at 1.5V (from 1.65V).

Windows 8.1 Pro64 booted right up, RealtempGT was showing 30 celsius at idle.
Started Prime95 blend test and Fluidmark, let the system test for 8 hours. Stable, temperatures were in the 60's for the CPU and in the 50's for the GPU's.

Last night, I let the system on as I was watching a movie before going to bed. Fell asleep and this morning found the computer at the lock screen. Checked the event log and noticed it blue screened 4 times between 4am and 6:30am.
Started Prime, immediate blue screen. I know cause I tested, uping the voltage to 1.4V makes it stable.
However, my questions is: why was it stable until last night and why the sudden change in behaviour? Do I have a bad chip?

 

[Burpo]

I'd check the power supply.

 

[terente]

I'd check the power supply.

It's an Enermax 1250W unit, do you think it went bad?

 

[escrow4]

I'd chuck the 480, its a power hog, and downgrade the overclock on the CPU to 1.3v or less (emphasis less). If that doesn't stabilise it, I'd look into the PSU.

 

[Zodiark1593]

If 1.4v is stable, why not leave it there if everything else is fine. Since upping the volts stabilizes the cpu, I don't think the psu is at fault as higher volts pulls more power. I would run it at 1.4v for awhile, and see how consistent it runs. Or back down the clockspeed a tad if temps are a concern. It sounds like your speed may be on the edge of stability there, backing down a bin or so may be enough to achieve stability.

 

[Galatian]

Perhaps a heat issue? I'm currently overclocking a i7-4790K and I couldn't get it stable at 4,7 GHz with 1,28V, but now that I applied Liquid Ultra (which reduced top temperatures by almost 10 degrees) I passed 24 hours Prime95 with 1,26V.

Odd I know, but perhaps it was just around the limit anyway and you tested in a colder room back then. Now that we have summer the ambient temperatures might have been enough to trip it over the unstable point.

 

[cytg111]

so .. bsod'ed 4 times while idle? Unless the heatsink has fallen off or ambient temps have risen 50 degrees i think we can rule out temps.
Sorry dude, sounds bad. Driver updates lately? How about running a live-linux distro off usb or cd/dvd and stress test from that?

edit : upping the voltage makes it stable u say? sure sounds like its the cpu

 

[Zodiark1593]

so .. bsod'ed 4 times while idle? Unless the heatsink has fallen off or ambient temps have risen 50 degrees i think we can rule out temps.
Sorry dude, sounds bad. Driver updates lately? How about running a live-linux distro off usb or cd/dvd and stress test from that?

edit : upping the voltage makes it stable u say? sure sounds like its the cpu

My thoughts exactly. I think the higher volts actually making the cpu stable rules out temps. Like I mentioned, he was probably barely stable to begin with. A slight voltage increase, or clock speed decrease may be the key so to speak.

That said, while cpus are guaranteed by Intel for stock speeds, it's usually a bit of roulette as for how much further you can go. Looks like you (the OP) got the short end of that, sorry to say.

 

[cytg111]

but it is ivey-e, it cannot be that old? Never heard of, or experienced, degradattion that fast.. and from nothing to 4 bsods in a row? that shit usually creeps up on you. Something is off.

 

[Zodiark1593]

but it is ivey-e, it cannot be that old? Never heard of, or experienced, degradattion that fast.. and from nothing to 4 bsods in a row? that shit usually creeps up on you. Something is off.

Problem with the mobo perhaps?

 

[jason166]

The fact that upping the voltage fixes the issue sounds to me like your chip may have degraded... 1.35 volts is a little on the high side for 22nm, not that others don't run it, but you are taking a chance and perhaps your luck in the chip lottery ran out in this case.

 

[Smoove910]

what happens if you leave the CPU voltage at 1.35 and increase RAM back to 1.65v?

 

[cytg111]

The fact that upping the voltage fixes the issue sounds to me like your chip may have degraded... 1.35 volts is a little on the high side for 22nm, not that others don't run it, but you are taking a chance and perhaps your luck in the chip lottery ran out in this case.

- degradation characteristics for different proces nodes may differ! What we are seeing here could be the top of the iceberg for intel 22nm.

 

[Erenhardt]

- degradation characteristics for different proces nodes may differ! What we are seeing here could be the top of the iceberg for intel 22nm.

Yea. Also, my 7870 did 1250MHz at stock voltage when I first plugged it in. After a couple of hours 1200MHz was unstable! I think degradation is massive in the first couple of hours

 

[Zodiark1593]

Yea. Also, my 7870 did 1250MHz at stock voltage when I first plugged it in. After a couple of hours 1200MHz was unstable! I think degradation is massive in the first couple of hours

Don't think I've ever heard of PC parts needing a break-in, let alone non-moving components. :\