Re: [hatari-devel] Hatari profiler updates and DSP cycle questions

[ Thread Index | Date Index | More lists.tuxfamily.org/hatari-devel Archives ]


The results are now even stranger - but I expected that because the original results didn't make sense for +(e-1) or +(e-1)*2.

Previously I saw:

2 cyc   (P:INT, X:INT, Y:INT)
5 cyc   (P:EXT, X:EXT, Y:INT)  (+3 osc)  
6 cyc   (P:EXT, X:EXT, Y:EXT)  (+4 osc)

Now I see:

2 cyc   (P:INT, X:INT, Y:INT)
6 cyc   (P:EXT, X:EXT, Y:INT)  (+4 osc)
8 cyc   (P:EXT, X:EXT, Y:EXT)  (+6 osc)


For a real Falcon it should be:

2 cyc   (P:INT, X:INT, Y:INT)
4 cyc   (P:EXT, X:EXT, Y:INT)  (+2 osc)
6 cyc   (P:EXT, X:EXT, Y:EXT)  (+4 osc)

So something funny is going on with the cycle sum for external references. It is almost as if the external penalty is being added twice somewhere, with a different formula?

D.


Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/