Well, I haven't been monitoring my temps, but I did roll back my ATI driver to the Omega version of the old 4.12 drivers (I think these are about a year old). Since then, no problems playing on small and standard maps to the finish. (The only reboot error I had was loading a late stage save file, 1987 ish. The wierd thing is that it wasn't accompanied by the extremely high fan revving the old crashes were causing. Clearing the games cache allowed the game to load properly on the next try). I need to try a larger map next.
Whether the game is improperly interacting with more recent graphics API features in a way that causes overheating (some of my serious error recovery messages in WinXP referenced failure due to an infinite loop being created, leading to the crash), or is just causing crashes for non heat related issues seems a minor point to me. It seems that the coding of the game itself, in interaction with more current drivers, is causing crashes and added stress on some people's systems.
Some people have had success in stopping the crashes by using tweak utilities to force PixelShading to the 1.1 standard, vs. the 2.0 API standard implemented in current drivers and hardware. I'm guessing that the driver I am using with success now probably is coded to the 1.1 standard, or at least a version earlier than the one causing the problems here.
Blaming people for not using the latest drivers seems to be a common excuse for game companies when people experience problems like this and in many cases using the latest drivers is the proper advice. Ironically, in this case, for many the serious problems can only be fixed by rolling back their drivers to a much earlier time frame. When Firaxis updates their own drivers to the most current ones, maybe they will then be able to get somewhere in fixing this problem.