OK, this makes sense now that nath4 pointed out that I had refresh disabled on one of them and not the other... :/ Now it becomes a comparison of a normal loop that runs quickly (i.e. doesn't do anything that changes screen, so not brought down to only run at 30 cycles per sec) vs a non-refresh loop (where screen updates are purposely repressed for up to half a second).
