Hmm, I get 100% CPU usage as well.
The pentium does have a few power down related instructions, but for obvious reasons, these tend to be priviledged instructions.
I think Windows will issue power down code when it sees CPU usage is low. (According to the task manager, and how often there is nothing to schedule other than the idle thread).
I believe the use of the Sleep Windows API call will generally decrease the CPU usage, if used properly. OP2 seems to make somewhat odd use of Sleep and threads though. I suspect what really happens is at least one thread is always running, and possibly stuck in some sort of busy loop (perhaps waiting for other threads, or waiting for enough time to elapse). At least while a game is active. The CPU usuage remains fairly low at the main menu and such. I'd never actually really noticed the high CPU usage before. The thing is, the game isn't actually that busy, so you can run lots of other things at the same time, and they'll all take whatever they need, and OP2 will happily run on what's left. But it does seem to consume whatever is left available. Not so good for battery use on a laptop.
I'd had thoughts of mucking with the Sleep calls in the network code in an attempt to reduce lag. It actually uses Sleep with a non zero parameter between checking network buffers. I was a little concerned about timing issues (maybe the game depended on that to run at the correct speed?), although I thought CPU usage might be more of a concern. But, since it's apparently already at 100%, I might as well go ahead and edit away.
Btw, this is an excellent reason why you should never code busy wait loops (among a number of other reasons). I remember some early version of RiddleHack that was posted on here had this issue. I promptly complained about the CPU usage on that one. If you need a timing delay loop, make sure it calls Sleep.
