What's the Tjmax for a 45nm CPU?

[sam1275]

Hello.
I want to know the maximum core temperature for a AMD processor, which is 45nm process.
I extensively searched internet, asked several PC forums, even emailed AMD directly, nothing worked.
The common believed 70C is not correct, that maybe the case max value.
Could anyone help me?
Thanks.

[wraper]

As vast majority of AMD CPUs has die soldered to the heat spreader, die temperature is almost the same as case temperature. Claiming 45nm process is not useful BTW, it has nothing to do with it.

[sam1275]

As vast majority of AMD CPUs has die soldered to the heat spreader, die temperature is almost the same as case temperature. Claiming 45nm process is not useful BTW, it has nothing to do with it.

Thanks for replying.
The two measured temperature are not even near, and that CPU ran over 90C at junction for a very long time without any problem.
I remember someone said Tjmax is lower on a smaller process, but I cannot confirm that.

[sam1275]

I searched some Intel document and found the Tjunction seems do not have much interference with process size, instead they are very different when package change, for example, a i3 2350M have 85C Tjmax on PGA, but 100C on BGA. How do package method influence the max die temperature?

[wraper]

EDIT: did not see your last post

I remember someone said Tjmax is lower on a smaller process, but I cannot confirm that.

Then why 14nm intel 7700k has 100^oC tj?

The two measured temperature are not even near,

Wondering how you measured it? Put thermocouple in between of the heatsink and CPU? Doubt that.

that CPU ran over 90C at junction for a very long time without any problem.

Does not mean that max temperature was not exceeded, and, most likely, your CPU was throttling.

Where is your process max temp correlation?
http://www.pantherproducts.co.uk/posts/cpu-maximum-temperatures/

[shteii01]

It depends on specific model.

Since your question does not specify the model, it has no answer.

[TheAmmoniacal]

Random example, Opteron 6128 has a Tj max of 70 C as you mentioned. Of course they might last a long time far above this.. Not difficult to find this information, there are datasheets for CPUs too.

[sam1275]

Wondering how you measured it? Put thermocouple in between of the heatsink and CPU? Doubt that.

The BIOS report 2 different temperatures, one is CPU temperature which is lower, another is CPU MOS area which is higher.

[sam1275]

Random example, Opteron 6128 has a Tj max of 70 C as you mentioned. Of course they might last a long time far above this.. Not difficult to find this information, there are datasheets for CPUs too.

Unfortunately I cannot find any datasheet for my CPU, some forum also suggest AMD have a poor documentation on processor(Unlike Intel's ARK).

[exe]

I remember someone said Tjmax is lower on a smaller process, but I cannot confirm that.

That is my intuition as well, although it's not confirmed by any datasheet. Still, it's an absolute maximum rating. I would assume smaller transistors are more prone to the (diffusion?) damage that accumulates over time. In my experience, overclocked and overheated CPUs and videocards tend to degrade (in terms of max freq) or even die after some months of abuse.

But, again, I don't have any reliability data and I prefer to keep my processors well below maximum ratings.

[wraper]

Wondering how you measured it? Put thermocouple in between of the heatsink and CPU? Doubt that.

The BIOS report 2 different temperatures, one is CPU temperature which is lower, another is CPU MOS area which is higher.

If there is temperature sensor under the CPU, it does not show any close to real CPU case temperature in the middle of the heatspreader.