Linpack 11 temps - when torture just isn't enough

[Ajay]

Must resist urge to rant about why linear no threshold is a terrible model for stochastic radiation hazards... :sneaky:

Well, high doses aren't stochastic anyway and that cuts to the point IDC was making about high voltage and temperature.

----------------------------------------------------

IMO, this thread should be re titled "CPU torture is pointless".

 

[Idontcare]

IMO, this thread should be re titled "CPU torture is pointless".

CPU torture is awesome, if your goal is to flesh out the weibull plot for determining lifetime reliability statistics based on accelerated lifetime testing

But yeah, if you are just looking to determine if your processor is stable at a given OC then you probably don't need to take it to TJmax just to see if it fails or not

 

[2is]

10w more is what I was able to measure via a kill-a-watt at the same voltage from 4GHz to 4.8GHz, obviously if I honed 4GHz down I'd be below shipping voltage at around 0.987v, but for the point of my inquiry we keep them the same.

So my question is if 1.2v is accepable shipping votage for Intel, how much of a difference is 3.8GHz vs 4.3GHz going to actually make when it comes to electron migration if the amp change is so small going from 4GHz to 4.8GHz?

10 watts @ 1.2v is over 8 additional amps going through those transistors. That's a respectable amount of current. 10 extra watts from a 120v outlet is less than .1 amps.

 

[BallaTheFeared]

10 watts at 1.27v

If we're to assume 4GHz was pulling 90w and 4.8GHz was pulling 100w..

70.87 AMPS for 4GHz

78.74 AMPS for 4.8GHz

Roughly an 11% increase in AMPS, I should care about this as far as shortening the lifespan on my processor?

I would think the real issue would be 95C or higher operation at 1.2v, not weather it was at 3.7 or 4.3. Voltage + Heat = Problems. I've never heard Voltage + Heat + Clock Speed = Problems before.

What strikes you as more problematic, 3.7GHz 1.2v 99C operation or 4.3GHz 1.2v 60C operation? In the past I would have said the former, do I need reeducation? I mean I guess 3.7GHz 1.2v 60C would win out here, but is that AMPERAGE something I need to concern myself with as far as what I should clock my CPU to?

I was under the impression moderate voltage with a decent clock speed while keepying the processor cool was ok, is it not? I have no context to work with, it's degrading faster isn't something I can run with when deciding if I want to run 4.9GHz 1.34v or 4.8GHz 1.27v or 4.6GHz 1.17v or 4.4GHz 1.1v...

What do I do with this information, how do I apply it to my situation? How can I weigh the pros and cons if I don't exactly know what the cons are? Is 11% more AMPS truely something I need to worry about? The context comes from Intel boosting VID for AVX code and getting quite close to his 4.3GHz OC as far as voltage goes....

I'm just not sure how to take this, I understand my processor is slowly dying even at stock, but if keeping my processor cool and running it with moderate voltage is greatly accelerating it's decline than I need to take that into consideration but I can't since I have no idea how much clock speed is affecting any of it, I know what acceptable voltage is, I know what decent temps are, I have no idea how to context clock speed in this.

For instance I see people running 4GHz 920s for years, four years now? A 4GHz 920 is going to be pulling 200w or more, at 1.3v or more, it's amperage is going to be far higher than haswell at the clocks/voltage we're discussing... Same with Core2, people running those at 4GHz for even longer, even more votage, even more power draw... I know their different uarch, different nodes, different materials... I feel I have no context from which to decide from, and that's scary to me.

 

[2is]

My personal opinion is that, no, I don't think it would adversely affect the useful life of the CPU, but I get where IDC is coming from with his technical explanation.

Obviously the processes of degradation is significantly increased with higher temps and/or volts.

 

[Ajay]

CPU torture is awesome, if your goal is to flesh out the weibull plot for determining lifetime reliability statistics based on accelerated lifetime testing

But yeah, if you are just looking to determine if your processor is stable at a given OC then you probably don't need to take it to TJmax just to see if it fails or not

I used to do allot of accelerated testing when I worked as an engineer in manufacturing (product development in plastics related companies). This included environmental chambers and designing jigs to apply repetitive stresses while cycling temp and humidity - fun times :thumbsup:

 

[Ferzerp]

My 4.4Ghz IB doesn't get over the low 60's in linpack. No killing here. Sorry.

 

[JimmiG]

My 4.4Ghz IB doesn't get over the low 60's in linpack. No killing here. Sorry.

That's because it doesn't have AVX2..

 

[BallaTheFeared]

That's because it doesn't have AVX2..

Attaboy JummiG, tell him it's because his is soooo slow ()