Would you consider this a safe i7 930 Overclock?

T

t1337Dude

Limp Gawd
Joined
Mar 29, 2010
Messages
329
Since my Gigabyte motherboard is extremely finnicky about recognizing memory with the right speeds, I had to overclock my i7 930 to 4.2ghz so all 6GB of my RAM would be recognized at DDR3 1600. It idles at 40c and I just played Crysis maxed for 15 minutes, alt tabbed out, and RealTemp said I was at 49c. Is this a safe overclock? Do you think I could push it further in the future if I improved my cooling? All I'm using is air cooling with a Hyper 212 Plus heatsink. Do you think the temperatures will go down after the Arctic Silver 5 paste "burns in", so to speak?
 
and RealTemp said I was at 49c.
Click to expand...

The second you alt tabbed the game stopped running the temps dropped. Core temps drop almost instantly when the load is removed. At least that is my guess because that is an incrediable temp and either you have the best cooling I have every seen or something is not right.

I guess its possible but bejebbus dont touch anything if it is.

Bring up realtemp and run something like OCCT or Prime 95 or Intel Burn Test at the same time and let us know what that says.

Chip is not even breaking a sweat until you get around 90C ish. Somewhere around 100C it will send out a signal to the motherboard to run the fans at 100%. Somewhere over 100C ish it will just shutdown to protect itself.

Your voltages are what you should be worrying about, thats where cpu damage (even if it is running cool) is most likely to occur.
 
BillParrish said:
The second you alt tabbed the game stopped running the temps dropped. Core temps drop almost instantly when the load is removed. At least that is my guess because that is an incrediable temp and either you have the best cooling I have every seen or something is not right.

I guess its possible but bejebbus dont touch anything if it is.

Bring up realtemp and run something like OCCT or Prime 95 or Intel Burn Test at the same time and let us know what that says.

Chip is not even breaking a sweat until you get around 90C ish. Somewhere around 100C it will send out a signal to the motherboard to run the fans at 100%. Somewhere over 100C ish it will just shutdown to protect itself.

Your voltages are what you should be worrying about, thats where cpu damage (even if it is running cool) is most likely to occur.
Click to expand...

I *think* I'm at 1.344v. And when I alt-tabbed out of Crysis, it was put into window mode and the game was still running behind RealTemps. I'll try running the Intel Burn test now.


EDIT: Ran it once and my temps jumped to about 70 average (between the 4 cores). Ran it twice and then it BSOD. Ran it a third time and it gave me another BSOD. I take it that this isn't a safe overclock then?



EDIT 2: I took my voltage down to 1.29v and so far it's working just fine. Since I'm a bit ignorant to the subject, what does voltage impact? All it does is sustain overclocks, correct? It doesn't improve performance or anything of the sort, right? And of course drive up temperatures and shorten the life of the CPU. Meaning that I want the lowest voltage possible to sustain my current clock speed?
 
Last edited:
It could be more stable, but it's probably still safe.

I'm satisfied if I can run OCCT CPU Linpack test.

5 iterations on the original test was said to be stable, but I'm satisfied with just 3. I don't think you could replicate the same load with any program... except for Core Damage, which is another CPU stress tester (said to run even hotter).

Edit: And yes - lowest stable voltage (sometimes I can bump this down in cold seasons, or turn the clock up a bit).
 
Thank you. I'm sort of disappointed because I wanted to overclock as time went on, instead of just "dropping the bomb" on my processor right when I got it. But I'll be damned if I get anything less than my advertised DDR3 1600. I've had slow RAM all my life and it's about time I sped it up. My Windows Index gave my RAM a 7.8, which I'm now satisfied with (was 7.4 earlier). Does anyone here know how high you can take an i7 930 with good cooling?
 
4.4-4.5 GHz is the absolute maximum for a 920/930 and you'd probably get that on water. You might be able to go a bit higher with extreme cooling but that wouldn't be something for day-to-day usage.

Actually, I've never seen anyone using anything over 4.4 24/7. I think I've seen a few running 4.3 but most run within the 4.0 to 4.2 range.
 
i7 920/930 D0 steppings vary in overclockability.

With high-end air, people tend to get between 4.0-4.5Ghz, but I've seen up to 4.7Ghz on air... not sure if that was 100% stable, because that's high for water too.

Your chip seems to be right in the middle (but then there are other variables: motherboard settings, the motherboard itself, and the fact that your heatsink could be bigger (don't know if that'd make much difference though)).
 
Concerning the voltage, yes, you are right on. Its a bit early and my major concern was that some motherobards will set some pretty high voltages under auto but once you get the major settings worked out it is a good practive to test and get the OC you want at the lowest voltages you can for the reasons you mentioned.

Damn I gotta get me one of these things. I just cant get over how well they do without major fussing around.
 
I typically run 20 runs of Linx which includes the same temp stress method as OCCT and Intel Burnin Test. If the processor passes and stays below 83c then I can move on to a longer term run. I like to use Real Temp because it tracks the min and max for a given time period for you. I typically reset the temps before each run. With my case being pretty small, long term temp stability is an issue. So after the LinX run I do a longer 8-24h run. If you can pass a 24h run more than one time (I typically do the first soon after the Linx run and then do a spot check a few days later when convenient). I typically notice temps rising higher in a 24h run than LinX because having a hot CPU running in a small case over time causes ambient temp to rise. With good case airflow this should be less of problem or decent water cooling (good external heat exhausting and fresh air intake).

When you have a chance can you please post your OC in the 930 OC'ing thread here. http://hardforum.com/showthread.php?t=1502931
 
Thanks guys. I can't even get sleep tonight because my brother keeps waking me up due to crashes from unstable voltages. Once I find a stable balance, I'll post in that thread. If my voltage happens to be too high but there are no noticeable crashes and the temperature is in a safe range, is there still a chance of damage to the processor?
 
t1337Dude said:
Thanks guys. I can't even get sleep tonight because my brother keeps waking me up due to crashes from unstable voltages. Once I find a stable balance, I'll post in that thread. If my voltage happens to be too high but there are no noticeable crashes and the temperature is in a safe range, is there still a chance of damage to the processor?
Click to expand...

You push your Voltage above the Intel recommended maximum at your own risk. You can kill your processor even if the temperatures are in a safe range. Think of it as a rubber bang with a given number of time you can stretch to power a sling shot. Your streching it longer so that you get more power / rang from your sling shot. The second you strech it beyond the manufacturer's spec he can't guarantee how long the rubber bad will last. Typically over streching means your rubber band won't last as long over time. But there is a limit that if you go beyond that the band will break right away. It's your call. There is no feasible way to identify what the breaking point is beyond breaking it. The mfg get's their data from breaking a statistical sample and projecting that on the expected life time of the part.
 
You must log in or register to reply here.
Back
Top