Reply to: My hard working Apple rMBP stays cool

[scooterscot]

It's called Xeon Phi, and it's already here. Of course you can't get them in macbooks, you need a proper computer.

" If it was released right now, it would occupy the third place in the top 500 list of supercomputers."

Wooah. Yes apple you must.

[doodab]

In my opinion intel should just start slapping more cmos cores into the package for performance. Imagine 32 cores under the hood in my mbp, I'll give it a couple of year.

It's called Xeon Phi, and it's already here. Of course you can't get them in macbooks, you need a proper computer.

[doodab]

Do you reckon there's sensor embedded on the chip itself? I would have thought surface mounted thermistor in the vicinity. Show's you how far things have come since me uni days.

There have been on-die temperature sensors (called DTS) for ages. There is a register that reports the number of steps which the temperature is below the factory calibrated "control" temperature, which is calibrated so that it reports 0 when the control temperature is reached, and that's when the thermal control circuit kicks in and starts reducing clock speed and voltage to avoid damage.

There is another temperature (130C I think) at which the chip just shuts down, which should only be reached in the event of a total cooling failure.

[scooterscot]

Having your CPU crap out on you is worse.

Never had it happen once in 23 years since my first 86dx.

In my opinion intel should just start slapping more cmos cores into the package for performance. Imagine 32 cores under the hood in my mbp, I'll give it a couple of year.

Last night my machine converted a 10gig mkv file in 55 minutes

[scooterscot]

It's a shame you're talking complete bollocks again.

Try this for some very basic information that even you may be able to grasp:
The Effect of Heat on CPUs and Computer Systems

Well the life of the component is still the same. You're talking about wear out. I'm not. I'm taking about capability. Operating the component at temperature is no big deal.

As intel says

"Ivy Bridge has a higher TjMAX specification, which governs when the CPU starts throttling in order to protect itself from heat damage. The cut-off for the Core i7-3770K is 105°C."

[d000hg]

Yes it is. Nothing worse that sweaty hands whilst you're typing.

Having your CPU crap out on you is worse.

[Ticktock]

Why? Because it's passed the boiling point of water?

I know IT folks don't do science but let me hit you with some anyway: Silicon has a melting point of 1,414 °C, It's not going to feel that much stress at 100.

Beside any material that experiences heat is stressed, which is only counteracted by it's strength. This deterministic understanding in isotropic materials is well understood.

It's a shame you're talking complete bollocks again.

Try this for some very basic information that even you may be able to grasp:
The Effect of Heat on CPUs and Computer Systems

[scooterscot]

The sensors in the chip measure the temp per individual core, or the averaged temp of the hottest core, depending on how you read it (i.e. which interface you use to get the numbers out). Your gauge could be showing the temperature at which throttling will start (Tjmax), as that is calibrated per chip I believe, and 102 is the right ballpark. For example my ancient Lenovo shows Tjmax = 105C, Core 0 = 39C and Core 1 = 31C.

Do you reckon there's sensor embedded on the chip itself? I would have thought surface mounted thermistor in the vicinity. Show's you how far things have come since me uni days.

[scooterscot]

It's a clever Apple design which keeps all the heat inside the case so your hands don't get warm.

Yes it is. Nothing worse that sweaty hands whilst you're typing.

[scooterscot]

When I'm maxxing out my GPUs I worry when the CPU hits 70c, I've never had go beyond that even overclocking it heavily

105c has got to be stressing the CPU

Why? Because it's passed the boiling point of water?

I know IT folks don't do science but let me hit you with some anyway: Silicon has a melting point of 1,414 °C, It's not going to feel that much stress at 100.

Beside any material that experiences heat is stressed, which is only counteracted by it's strength. This deterministic understanding in isotropic materials is well understood.

[amcdonald]

When I'm maxxing out my GPUs I worry when the CPU hits 70c, I've never had go beyond that even overclocking it heavily

105c has got to be stressing the CPU

[d000hg]

102C is hot. That's around the temperature the chip will start to throttle itself to avoid dying.

It's a clever Apple design which keeps all the heat inside the case so your hands don't get warm.

[doodab]

Or it could be just cumulative temperature, it does have several cores after all.

The sensors in the chip measure the temp per individual core, or the averaged temp of the hottest core, depending on how you read it (i.e. which interface you use to get the numbers out). Your gauge could be showing the temperature at which throttling will start (Tjmax), as that is calibrated per chip I believe, and 102 is the right ballpark. For example my ancient Lenovo shows Tjmax = 105C, Core 0 = 39C and Core 1 = 31C.

[scooterscot]

102C is hot. That's around the temperature the chip will start to throttle itself to avoid dying.

Or it could be just cumulative temperature, it does have several cores after all.

[doodab]

102C is hot. That's around the temperature the chip will start to throttle itself to avoid dying.