How does this setup look for everyone on air.

[WFO]

Wow! I've never followed WCG threads before. You guys seen ultra conservative on temps. I remember when I-7s first came out, the Intel OC threads indicated 85C was normal for a good clock on air. Many said it was 24/7 safe. Did something change? Lots of dead chips? RealTemp indicates a 100C TJ Max. OC guides suggest anything under that is safe.

While it's apples and oranges I used to run my AMD laptops in 106F heat while folding. The chips reached 73-74C with no damage. I wouldn't want to run my X6 that hot.

Did I miss some required reading somewhere? I generally go for the highest stable clock prior to the one requiring a huge v-core increase. That's just me.

[D_A]

WCG requires stability like nothing most overclockers can dream about. These machines aren't just run 24/7, they're 24/7/365 at 100% on all cores. Running at temperatures that some consider "safe" will end up cooking not only your chip but your board as well. Remember, more heat means less lifespan for any semiconductors. The further you can keep away from the limits the longer your gear will last.
We've had people come here, one after the other, claiming their machines were stable by all the "standard" markers (prime, linX etc), but they were producing errors on WCG workloads. Almost every time they needed to drop their clocks a couple hundred MHz before they were stable enough.

[WFO]

WCG requires stability like nothing most overclockers can dream about. These machines aren't just run 24/7, they're 24/7/365 at 100% on all cores. Running at temperatures that some consider "safe" will end up cooking not only your chip but your board as well. Remember, more heat means less lifespan for any semiconductors. The further you can keep away from the limits the longer your gear will last.
We've had people come here, one after the other, claiming their machines were stable by all the "standard" markers (prime, linX etc), but they were producing errors on WCG workloads. Almost every time they needed to drop their clocks a couple hundred MHz before they were stable enough.

In the 5 days I've been here, I haven't had to reduce clocks. I did have to increase v-core 1 step on 2 rigs. Since then everything has been 100% stable and also have 100% on return rate. Oddly 2 GTX 295s could only fold on 1 GPU. With GPUgrid they happily crunch on both.

I had to ask. IntelBurn Test and Folding stable hasn't been WCG stable but was easily corrected. I usually only look for a 2 year lifespan. After that, work produced per KW is not worth running the rig. I'm a CA user subject to the Commie Cap and tax power rates.

Edit: That's not political. It's the truth.

[D_A]

No, calling anything "commie" is political and a pointless discussion I will be no further drawn into.

[CrazyJerry]

4.81GHz, 4.18Ghz

For long term sub 70c is considered safe, Zalbard's advise is good. Even though it seems like we should push everything we can get out of a CPU that extra 100mhz really is not worth it sometimes vs being 100% stable.

Yes 4.18...lol didn't even see that typo... or I was just dreaming!

[CrazyJerry]

WCG requires stability like nothing most overclockers can dream about. These machines aren't just run 24/7, they're 24/7/365 at 100% on all cores. Running at temperatures that some consider "safe" will end up cooking not only your chip but your board as well. Remember, more heat means less lifespan for any semiconductors. The further you can keep away from the limits the longer your gear will last.
We've had people come here, one after the other, claiming their machines were stable by all the "standard" markers (prime, linX etc), but they were producing errors on WCG workloads. Almost every time they needed to drop their clocks a couple hundred MHz before they were stable enough.

Noted...

I do appreciate the feedback and on this thread alone. The knowledge I have gained has been great. I was just wanting to learn more about the i7's and what they can do. I also want my gear to last as long as possible. Once I get her tunned in she'll run until the day a component fails. Hopefully not the MoBo or CPU.

I know most people won't like this next comment...
I just wanted to get the highest stable OC to run WCG and do as much as possible for this week. I plan on toning it down a bit so that I have 100% stability across the board. I can't have this system crash or lockup. My company said I could build a system for home work use. 30% of what I do is night and weekend work remotely. Local work is testing Virtual Machines for our production environment, large database conversions and HD video editing.
95% of the time I will be running WCG.

That said Thanks all for the help and feedback. Current temps avg. of 60 deg. @3.9ghz

[D_A]

I know most people won't like this next comment...
I just wanted to get the highest stable OC to run WCG and do as much as possible for this week. I plan on toning it down a bit so that I have 100% stability across the board. I can't have this system crash or lockup. My company said I could build a system for home work use. 30% of what I do is night and weekend work remotely. Local work is testing Virtual Machines for our production environment, large database conversions and HD video editing.
95% of the time I will be running WCG.

That said Thanks all for the help and feedback. Current temps avg. of 60 deg. @3.9ghz

There's a unwritten motto around here that sums things up nicely.

Run what ya brung

In other words, every contribution is valued whether it's from a SR-2 based system clocked to the point of creating a data singularity or a one-a-day return from an old Pentium2, it all helps the cause.

[artemm]

In other words, every contribution is valued whether it's from a SR-2 based system clocked to the point of creating a data singularity or a one-a-day return from an old Pentium2, it all helps the cause.

QFT, though

[CrazyJerry]

There's a unwritten motto around here that sums things up nicely.



In other words, every contribution is valued whether it's from a SR-2 based system clocked to the point of creating a data singularity or a one-a-day return from an old Pentium2, it all helps the cause.

:-)