I have read that there is really no such thing as real idle time since the CPU calculates idle time using idle loops, therefore the CPU is always at 100% use.

So what does this mean? It doesn't matter at all leaving the computer alone with nothing to do than playing a super-resource-intensive game???

I'm saying this because I wish to use my computer for Distributed Computing and was wondering that if idle time really does not exist, then at least use it for a good cause rather than for idle loops. My computer only uses about 10% of its CPU normally, so I was planning to dedicate another 10% for DC. So does this mean I can dedicate 90% for DC and it wouldn't matter since it would be used for idle loops anyway?

Any expertise on the subject?

I'm no expert in this but a simple proof that what you are saying is wrong is the heat and fan speed. You're CPU is cooler when idle and the fan is slower so... there is a difference

This is a very good question. I've always wondered the same thing. On top of the idle loop, the processor is always polling for I/O data to process from the hardware. But obviously the idle processing isn't nearly as intensive as hardcore gameplay (or Excel ).

Calculating idle time is very gentle hardware wise, it's saying "I've got this much free because I've got nothing better to do." Most of the distributed computing applications have settings that say "only utilize x%age of the idle time" or "only when idle" in some form or another.

Actually, the computer does not "loop" to calculate idle-time.

When the dispatcher finds that it has no work to do, it stores the current time-of-day clock value and then (conceptually...) halts the CPU. When it's ready to resume doing work, it captures the new time-of-day value and subtracts it from the previously stored value. The difference between the two is the amount of time the CPU has been idle, and this is added to idle-time accumulator.

It is entirely normal for any given CPU to spend most of its time "idle," because most things are somehow constrained by the speed of some physical device. Microseconds add up fast.

The actual processing is somewhat more complicated, especially in multi-core/multi-CPU situations, but this simple explanation will suffice.