ODwilly wrote:My in-progress Prescott system tells me to contact HP to update the cpu microcode when it starts up and that the cpu is unknown. Yet the bios detects it as a Intel Pentium 4 670, which it is. Seems to work good so far, if it runs stable and does not give you any problems who cares right?
At least it lets it run. Some boards are complete bungholes about this.
I investigated the voltage regulator on an IBM 440BX board, this was one of their fancy "Intellistation" workstations that people paid a ton of money for when new. The kind of thing you'd think would get good long term support.
I found that the VRM was Coppermine compatible. Great, so I plugged in a basic 600E Coppermine CPU. It would boot, but then the BIOS would pop up a message saying the current CPU is "unsupported", and after acknowledging the message it would forcibly shut off. "Yeah this CPU works, but we don't want you to use it so I'm turning off now. Bye." No BIOS update available to fix it, it arbitrarily refused to go beyond Katmai.
Had an equally frustrating problem with an HP of the period, but that involved it refusing to recognize memory that was compatible with the chipset, but didn't exist in the marketplace when the computer was new. Again, that was a formerly high end expensive workstation, where you'd expect more support than they gave.
Hacking the BIOS on machines like this would be an interesting thing to try, but what kills me about these IBM/HP/Intel/Dell etc boards is that they always have soldered flash chips. Why do cheap consumer boards have sockets, but almost every expensive workstation/server board I've seen has the chip permanently soldered? I don't get that, at all. You'd think the more expensive the board, the more serviceable they'd want it to be. If not for this, I could get adventuresome with a flash programmer.