There’s been some occasional flareups about Intel using clock-throttling on Willamette processors.
Well, we mentioned this briefly a little while back, but AMD CPUs have a form of clock-throttling in place, too, actually pretty similiar to Intel’s. They even use the
From AMD Athlon Processor Model 4 Data Sheet, pages 10-11.
Let’s look at the key phrase:
The Southbridge can force STPCLK# assertion for throttling to protect the processor from exceeding its maximum case
temperature. This task is accomplished by asserting the THERM# input to the Southbridge. Throttling asserts STPCLK# for a percentage of a predefined throttling period.
When the CPU gets too hot, a motherboard can tell the CPU to stop working part of the time until it cools down.
There is nothing bad about this kind of clock-throttling in-and-of-itself, unless you like CPU keychains.
What could be bad about a particular clock-throttling situation is if a processor is rated at XXXXMhz, but cannot regularly run at that speed without these slowing down mechanisms in play.
There’s been periodic assertions that that is the case with Willamette, but at least some of the evidence presented has been later retracted by their originators.
Is Clock-Throttling A Regular Feature of AMD Motherboards?
I don’t honestly know for sure; it’s something the motherboard is supposed to do under dire circumstances. Palominos will have this built into the CPU, you will get this triggering no matter what then. At what point does this trigger? We don’t know, AMD doesn’t say.
The impression I’ve gotten from some of my heat travails, though, is that we’re talking pretty dire, somewhere over 70C, maybe a good deal over 70C.
So we’re talking about either one big damn fool with a hardy chip and a lousy heatsink or one miserable mofo mobo sensor before this gets triggered.
There’s also a second kind of throttling talked about in the AMD piece, the part talking about changing ACPI registers?
I guess you’d call that “voluntary” throttling. Those cooling tweaks, whether for Via or AMD boards? That’s how they work; they enable the second kind of clock-throttling.
I bet a few people who denounce Intel for clock-throttling are doing the same thing to themselves without even knowing it. 🙂