[Mahn]

And then you realize you could actually take it a step further and before you know it you are hunting for documentation on the amount of cycles each intruction and data access takes and other obscure clocking info, nevermind there are a thousand and one emulators already in existence that are several orders of magnitude better than the half assed attempt you are trying to come up with. Fun times :)

[tptacek]

That's such a huge problem with exploratory programming: the demotivating effect of knowing that there are better versions of almost anything you're building, because other people have been working on the problem for longer.

You have to get over it and keep going anyways or you'll never become one of those people yourself! (Or even be able to make an informed decision about whether you want to).

Not coincidentally: part of the point of the company we just started. :)

[goalieca]

Cycle accurate emulation of modern x86 would be impossible. Intel simply won't release the secret sauce on branch prediction, cache prediction, out of order execution, internal instruction engine, etc.

[im2w1l]

On the other hand, the absence of such information means that games don't depend on it.