Re: [hatari-devel] Correct cycles values?

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


 Hi,

Am Fri, 15 Mar 2013 19:52:23 +0200
schrieb Eero Tamminen <oak@xxxxxxxxxxxxxx>:

[...]
> 
> Ok, what about TAS and MOVE.W instructions, can they take several
> hundred cycles?
> 
> WARNING: cycles 272 > 256:
> $e12bc6 : move.w    d7,(a5)                    0.00% (14, 3536, 5)
> WARNING: cycles 284 > 256:
> $e12bcc : tas       (a5)                       0.00% (66, 14804, 0)
> WARNING: cycles 284 > 256:
> $e12bcc : tas       (a5)                       0.00% (67, 15088, 0)
> WARNING: cycles 284 > 256:
> $e12bcc : tas       (a5)                       0.00% (68, 15372, 0)
> WARNING: cycles 272 > 256:
> $e12bc6 : move.w    d7,(a5)                    0.00% (15, 3808, 6)
> WARNING: cycles 284 > 256:
> $e12bcc : tas       (a5)                       0.00% (70, 15676, 0)
> 
> The context for these instructions is:
> $e12bbe : addx.w    d6,d6                      0.00% (1, 4, 1)
> $e12bc0 : move.b    $be(a6,d6.w),$ffff(a5)     0.00% (1, 16, 0)
> $e12bc6 : move.w    d7,(a5)                    0.00% (1, 272, 1)
> $e12bc8 : adda.w    a2,a0                      0.00% (1, 4, 0)
> $e12bca : adda.w    a3,a1                      0.00% (1, 8, 1)
> $e12bcc : tas       (a5)                       0.00% (17, 4340, 0)
> $e12bce : nop                                  0.00% (17, 68, 1)

where does a5 point to? Could it be that TOS tries to start the blitter
here? The blitter then might "steal" all those bus cycles so that it
appears that the instruction itself takes very long.

 Thomas



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