Re: A89: TIGCCLIB 2.0 released!
[Prev][Next][Index][Thread]
Re: A89: TIGCCLIB 2.0 released!
On Sat, May 27, 2000 at 15:42:19 -0400, Scott Noveck wrote:
>
> That being said, let me point out one chunk of code from the grayscale
> portion of nostub.h:
[snip, pretty printed below]
> That loop is currently using up 1,442 clock cycles and being called 87.5
> times per second. That's 126,179 clock cycles per second - MASSIVE speed
> loss.
(Assuming no wait states:)
moveq #79,d0 ; 4
\loop
movem.l (a0)+,d1-d7/a2-a6 ; 12+8*12
movem.l d1-d7/a2-a6,(a1) ; 8+8*12
lea 48(a1),a1 ; 8
dbra d0,\loop ; 10/14 (taken/not taken)
4 + 80*(12+8*12 + 8+8*12 + 8) + 79*10 + 14 = 18,408 clock cycles ?!?
How did you end up with 1,442 clock cycles?
/Johan
PS. Scott, I always wanted to do a simple test on a HW2 calc and perhaps you
can help me. Use a stopwatch and measure the (exact) time it takes for the
calc to APD. By knowing this time, I can easily calculate the speed of the
oscillator that controls the LCD and the timer interrupts. I have reason to
believe that this oscillator is slower on HW2's. Thanks in advance... DS.
References: