Tim, This is on the Fractint bug list: Try cycling between float and integer math.

...

...

...

...

16/32 bits problem ? : With this in a PAR file:

squares { ; Floating Point !! reset=2000 type=formula formulafile=thomas.frm formulaname=Squares passes=t center-mag=-1.64e-006/2.205e-006/0.6666676 params=3/0 float=y cyclerange=0/253 sound=beep/fm/quant colors=@default.map } squares2 { ; Floating Point !! reset=2000 type=formula formulafile=thomas.frm formulaname=Squares passes=t center-mag=0/0/0.001 params=3/0 float=y cyclerange=0/253 sound=beep/fm/quant colors=@default.map } and that in THOMAS.FRM : frm:Squares { s2=sqrt(2) ampl=1/0.02 ang=3.1415926535898/real(p1) cs=cos(ang) sn=sin(ang) px=pixel a=real(px) b=imag(px) z=0: c=s2*(a*cs-b*sn) d=s2*(a*sn+b*cs) a=c b=d z=z+1 ampl*(abs(a)+abs(b))>1 } Between my old 486 and my Celeron (out of windows) there is not at all the same result !!Why ? (I have images if you want) Limit problem: With the same formula,switch "Floating point" to off. Edit the formula and change 1/0.02 by 50 In the first case, maxiter is near 50 (I don't know exactly) and in the other one,only 15 !! <<<<<< From what I can determine, the conversion to the float parser formula gets messed up. The value in vsp and LastInitOp are not correct when running with floating point math. Jonathan