Hi,
On torstai 14 maaliskuu 2013, Nicolas Pomarède wrote:
Le 14/03/2013 11:01, Eero Tamminen a écrit :
Shouldn't the cycles shown in profile disassembly then be:
cycles<< nCpuFreqShift
?
Yes, if you want to display cycles in a human readable way,
nCpuFreqShift should betaken into account.
Ok, I commited that change. The bootup times visible to eye now
about correspond to ones calculated from spent cycles.
Do these cycle values look about correct for 16Mhz Falcon:
$e00030 : move #$2700,sr 0.00% (1, 20, 1)
$e00034 : move.w $ffff8006.w,d0 0.00% (1, 12, 1)
$e00038 : reset 0.00% (1, 520, 1)
$e0003a : move.w $ffff8006.w,d0 0.00% (1, 12, 1)
$e0003e : move.w #7,$ffff8940.w 0.00% (1, 16, 1)
$e00044 : cmpi.l #$fa52235f,$fa0000 0.00% (1, 24, 1)
$e0004e : bne.s $e0005a 0.00% (1, 12, 1)
[...]
$e0005a : move.l #$808,d0 0.00% (1, 12, 1)
$e00060 : movec d0,cacr 0.00% (1, 16, 1)
$e00064 : moveq #0,d0 0.00% (1, 4, 3)
$e00066 : movec d0,vbr 0.00% (1, 8, 1)
$e0006a : nop 0.00% (1, 4, 1)
[...]
$e00088 : move.w #$10,$ffff8282.w 0.00% (1, 16, 1)
$e0008e : ori.b #$25,$ffff8007.w 0.00% (1, 20, 1)
$e00094 : btst #6,$ffff8007.w 0.00% (1, 16, 1)
$e0009a : beq.s $e000c8 0.00% (1, 12, 1)
[...]
$e000c8 : lea $ffff8800.w,a0 0.00% (1, 8, 1)
$e000cc : move.b #7,(a0) 0.00% (1, 20, 1)
?