1.4+ Vcore on a Gulftown. Should I worry?

[Mpegger]

Before anyone answers the title, I am not actually running my 980x @ 1.4+ Vcore. In fact, the Vcore is currently set at 1.28125, with a Dynamic Vcore of +.04375. For those who don't want to do the math, that's 1.325 Vcore under load, being reported at 1.312 Vcore via sensor.

Anyway, I've been playing around with the multi on this CPU and have it currently running with a 30x multi (stock BClock @ 133.2xx) at 1.325 Vcore (10hrs OCCT test). On this particular board, I also have Turbo enabled which allows all the cores to go up to a 31x multi, as well as the advanced CPU states (C3, C6, C7), which allows a single core to go a further multi, to 32x when the other cores are idle.

OCCT loads all the cores at 100%, and for the entire duration of the test, the Vcore was reported at 1.31 (31x multi). However, after running OCCT, I fired up my single thread test to make sure the CPU would still be stable with a single core at the 32x multi, which it seems to be (the test only last less then 9 minutes, but I ran it multiple times not locked to any single core). However, during the test, a couple of times the Vcore jumped up to 1.4+ Vcore, one time even being reported as 1.5(!!) Vcore. OCCT only polls the sensor once a second, so I do not really know how many times the Vcore may have jumped up, but it only last for one of the poll intervals, and only happens (reported) 2-3 times during that 9 minute test (although there were a number of jumps that stayed within the 1.3x Vcore range).

So, the question of this thread, should I worry at all about what seems to be occasionally high Vcore on a multi that won't be accessed on a regular basis? Is it just some random incorrect reading, though it didn't happen at all with the 31x multi? Is there a program available that can poll the sensor at a higher interval so I can watch the voltage closely? Anyone else experience such voltage jumps when overclocking? Maybe VDroop setting is to blame?

Please, veteran and extreme sages of overclocking, sprinkle some of your wisdom on this thread.

[edit] After a reboot, I've had OCCT monitoring in the background. Only normal usage so far (no gaming yet) and already 1.39 Vcore is being reported for the max Vcore seen. [/edit]

[edit] 1.46 max reported Vcore now. I guess the 32x multi gets used more then I thought. [/edit]

 

[djscrew]

*shrug*

seems okay considering how brief it is...

some 2-3 steps of voltage fluctuation is normal, added the increase vcore required for a 32 multi it doesn't seem surprising. I doubt it will harm your chip unless it's also running super hot

 

[Mpegger]

Temps don't seem to be an issue. Highest so far reported at 78F room temp during testing was 61C (CPU is watercooled and very often idles at room temp).

Reading some posts here and there, it seems that Intel's max voltage for Gulftowns is around 1.55V, with up to 1.6V instantaneous when switching voltages. If this is actually going up to 1.5V for a eye blink, it shouldn't be a problem at all.

 

[freeagent]

1.4v is no problem. My 970 is a bad ocer, 1.4v is where I have to start to get anywhere

Yes its that bad.

I am ok with giveing it 1.5v too

Of course you probably dont need that much.. as 980s are supposed to be nice lol

 

[Mpegger]

The voltages it's set at isn't a concern since it's well within the safe range. Even the voltages you state are ok. My concern is the fact that it's jumping nearly .3V, from 1.312 (which it's set at) to 1.6 (which it shouldn't be going to), in the first place. Imagine your chip jumping from 1.5V to 1.8V. Even for a brief moment (couple of cpu cycles), I'm sure 1.8V can shorten the life significantly.

 

[freeagent]

The voltages it's set at isn't a concern since it's well within the safe range. Even the voltages you state are ok. My concern is the fact that it's jumping nearly .3V, from 1.312 (which it's set at) to 1.6 (which it shouldn't be going to), in the first place. Imagine your chip jumping from 1.5V to 1.8V. Even for a brief moment (couple of cpu cycles), I'm sure 1.8V can shorten the life significantly.

I had a UD5, that was my first 1366 board. I do not recall ever seeing the voltage scale like that. Are you useing the newest bios? Maybe turn off the dynamic vcore, mainly because Im not sure what it is Turn llc on, and set your voltage, and you should be good to go from there. Altho my old giga did droop alot under load, about .05v, maybe a touch more.

 

[Mpegger]

Dynamic Vcore allows you to raise the stock Vcore, without locking the Vcore, so that when the CPU is idle, it can still dynamically drop the Vcore to save some juice. LLC is on and set at the max (2) setting. It could be the reason for the problem, though I haven't tested it and I do not understand how this particular Bios/Board works it. And it is the latest Bios in use. Its the only way I'm able to access the 2nd Turbo multi when a single core is being taxed.