fix: rename Rewind_run_frame to run_frame and fix fast-forward logic

[Helaas]

Fixes the multiplied fast forward speed as reported in issue #677

[DrFlarp]

I've been thinking about the draw elision logic at minarch.c:5730. Instead of spreading the core.run calls evenly over t/n milliseconds, would it be better to race to get all n frames out ASAP and then do one longer wait call?

After all, frame pacing doesn't matter when it's all within a single vblank haha

Or revisit the elision line to something other than hardcoding 10ms there since it was based on a platform that we don't target?

[DrFlarp]

My concerns appear to be unfounded, this works great on all my testing scenarios with multiple workloads and FF speeds :)