| Re: [hatari-devel] WinUAE CPU core zero cycles issue |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/hatari-devel Archives
]
Hi,
On torstai 14 maaliskuu 2013, Laurent Sallafranque wrote:
> >Laurent, is it OK if I add OpcodeFamily updating to 20 & 21 cores,
> >i.e. for the default Falcon WinUAE CPU core?
>
> Of course it is OK, don't hesitate.
Ok, done. I hope this is fine:
http://hg.tuxfamily.org/mercurialroot/hatari/hatari/rev/8b344d863ff3
I'm still seeing zero cycles for Falcon emulation with WinUAE CPU core
though, for example for "move.b d0,d1" here:
$e0429a : lea $ffff8606.w,a6 0.00% (23, 104, 3)
$e0429e : st $1690.w 0.00% (23, 288, 3)
$e042a2 : tst.w $043e.w 0.00% (23, 196, 3)
$e042a6 : bne.s $e04320 0.00% (23, 184, 23)
$e042a8 : move.l $0466.w,d0 0.00% (23, 368, 23)
$e042ac : move.b d0,d1 0.00% (23, 12, 3)
$e042ae : and.b #7,d1 0.00% (23, 104, 3)
$e042b2 : bne.s $e04320 0.00% (23, 268, 23)
$e042b4 : move.w #$80,(a6) 0.00% (2, 32, 2)
Not all of the executions for it had zero cycles, but most did.
How the number of cycles can change for it?
- Eero
PS. To find such occurrences, just set DEBUG define to 1 in
beginning of src/debug/profilecpu.c, recompile, and enable
profiling e.g. for TOS v4 boot.