X5667 ES Q3QR @ LN2 # 3rd Vantage on HWBot

[Sam_oslo]

What are you giving yours for 24/7? I'm keeping it under 1.34v as a self imposed limit right now, more later when I get some balls

I've chickened out at VVT=1.32v and this give max 24/7 stable @3.8Ghz(172x22) 1.21v. Next option in my BIOS is VVT=1.34v and I'm really tempted to use it, but trying to learn more bout these chips, and see specially how much VVT they can take for 24/7 use, before I raise the VVT.

1.65v-1.7v on AC 24/7 no problem with memory 2000 MHz CL7.

@SoF

According to xoqolatl I like high voltages

I belie VVT~=1.5v will get me up to @4.4GHz and I will be happy with that for 24/7 use, and my RAM can do 2000MHz CL9, but not CL7.

Can you please explain how 2000MHz CL7 can make those high VVT safe for 24/7?

[SoF]

CL7 2000 can sometimes need past 1.5 vtt while CL8 same speed 1.35V is all you need - at least on my RIIE it is that way CL8 and CL9 doesn't make a big difference

[Sam_oslo]

CL7 2000 can sometimes need past 1.5 vtt while CL8 same speed 1.35V is all you need - at least on my RIIE it is that way CL8 and CL9 doesn't make a big difference

Yeah, but does any of these make High VVT safer for 24/7?

The x20 Uncore of my X5650 needs a lot of VVT, almost independent of the Core-multi or RAM-timings. My concern is, what's the max-safe VVT for 24/7 without killing or degrading these 32nm CPUs?.

[Xtreme Addict]

I've chickened out at VVT=1.32v and this give max 24/7 stable @3.8Ghz(172x22) 1.21v. Next option in my BIOS is VVT=1.34v and I'm really tempted to use it, but trying to learn more bout these chips, and see specially how much VVT they can take for 24/7 use, before I raise the VVT.



I belie VVT~=1.5v will get me up to @4.4GHz and I will be happy with that for 24/7 use, and my RAM can do 2000MHz CL9, but not CL7.

Can you please explain how 2000MHz CL7 can make those high VVT safe for 24/7?

I never bother about safe voltages with mem and vtt/qpi, I like good results and sometimes I have to help components raising voltage a bit

I stopped thinking about all those theories about relation vtt - vmem like 0.5v etc. After launchtime of first i7 I bought i7 965XE retail and used it with R2E and 3x1 GB Goodram PC10600 D9 GTR, old good microns. On aircooling I put 1.8v qpi (max stable on 965XE on air cooling) and 2.5v mem (those microns loved voltage). I was working 24/7 on this setup during a month, on those voltages, Guess what was dead after a month? R2E but I RMA'
ed it. CPU and memory are working till now, they didn't feel a degradation factor and so on. I killed about 30 E8xxx cpus but I never killed i7. And I just don't believe in too high vtt and too high vmem.

[Sam_oslo]

I never bother about safe voltages with mem and vtt/qpi, I like good results and sometimes I have to help components raising voltage a bit

I stopped thinking about all those theories about relation vtt - vmem like 0.5v etc. After launchtime of first i7 I bought i7 965XE retail and used it with R2E and 3x1 GB Goodram PC10600 D9 GTR, old good microns. On aircooling I put 1.8v qpi (max stable on 965XE on air cooling) and 2.5v mem (those microns loved voltage). I was working 24/7 on this setup during a month, on those voltages, Guess what was dead after a month? R2E but I RMA'
ed it. CPU and memory are working till now, they didn't feel a degradation factor and so on. I killed about 30 E8xxx cpus but I never killed i7. And I just don't believe in too high vtt and too high vmem.

I know what you mean. i7 45nm have been good at taking torture, including high VVT for 24/7. They have been almost indestructible. I'm not sure if the same applies to these new 32nm CPUs, but I hope they are as resilient too.

These kind of tests can give some indications too, even tho these are not the same as for 24/7. It is promising that this chip could take 1.65v-1.7v VVT in this test But I'm still a kind of chicken, and need to hear more tales like this.