CPU Showdown for Pancakes

If you ask people how they rate as a driver, most of them will say they are better than average. At first, that seems improbable until you realize one thing: people judge themselves by different criteria. So Sally thinks she’s a good driver because she goes fast. Tom’s never had a wreck. Alice never gets lost. You can see the same effect with CPUs. Some are faster or have more memory bandwidth or more instruction issues per cycle. But [Andrew] and [Scharon] at Tom’s Hardware wanted to do the real test of a CPU. How well can it cook pancakes? If you want to know, see the video below.

While your CPU might be great for playing video games, it has a surprisingly small cooking surface, so the guys needed a very small pan. The pan had grooves in it, so they slathered it with thermal grease. We doubt that’s food-grade grease, either.

With all the trouble we take to move heat away from the CPUs, you’d think they’d cook faster. The chefs set up the BIOS on an Intel CPU to turn thermal throttling to 105C. Even so, it took 16 minutes to cook a cake while running a stress program.

How would AMD fare? Apparently, not so well. While the Threadripper CPU had a larger cooktop, something made it shut down when it got too hot and so, the pancake barely cooked. Granted, the Intel chip shut down once, but with some help from an external fan, it finished the task.

We prefer pancakes on a CNC machine, honestly. If you search for pancakes, you’ll be surprised that there are many similar projects. But none that used a CPU as a cooktop. Maybe we should try with a Raspberry Pi 4.

from Hackaday https://ift.tt/326MZcL
via IFTTT