I know what a hard fault is: a piece of code is not in RAM and must be retrieved from virtual. What I want to know is: how many hard faults per second are too many?
I Googled until my fingertips were bruised. I could fill an entire hard drive with the same explanations of what a hard fault is, over and over in varying levels of detail. Even Microsoft’s Technet does it. In fact, they even say “A consistently high number of hard faults per second indicates” problems. But they don’t say how many is too many. What is “consistently high”?
It seems like the kind of thing that might vary between 32 bit and 64 bit operating systems, but not from one system to the next. The number of hard faults per second that I see will vary greatly from system to the next, but the number that is “too much” should be pretty consistent. I imagine that a 64 bit PC with 2 Gb of RAM is going to have a lot of them and the same PC with 128Gb will have very few. And the former is going to perform worse than the latter. But the actual number per second seems like it would be pretty directly related to the user experience of system performance.
So how many is too many?
