Flaming Durons, part deux
Jeff Tom over at AMD Zone wrote in to tell me that he'd put the color-changing theory to the test. He was in the unenviable position of having a Duron with a crunched core, likely due to a "heatsink incident." The core wouldn't run, but still generated plenty of heat, so it was perfect for this test.
Basically, he stuck a temp probe on the core, fired it up, and waited with bated breath for it to turn purple. He was disappointed, though not at all surprised, when the color stayed the same green it always was, even as the temperature topped out at nearly two hundred degrees Farenheit.
After confirming that heat wasn't responsible for the color changes, Jeff came up with some other possible explanations for the purple Duron syndrome, all of which sound a lot more plausible than the "heat tempering" argument. Go check it out.