VW9 performance slower than VW8.5 on boolean operations between solids

[Paolo]

Hello, I'm Paolo from Italy,

Using VW9 I realized that boolean operations between solids are very much worse than VW8.5.The evidence jump to eyes using boolean subtraction when one of the solids involved contains round parts (ovals or round rect) and is obtained by multiple extrusion.

Perform the following test in VW8.5 and VW9.0-------------------procedure test;varbody, hole, edge, difference : HANDLE;result : INTEGER;chrono : LONGINT;

beginchrono := GetTickCount;beginXtrd(0,10); RRECT(-20,20,20,-20,14,14);endXtrd;body := LNewObj;Move3DObj(body,0,0,-1);

beginXtrd(2,10); RRECT(-18,18,18,-18,12,12);endXtrd;hole := LNewObj;

beginMXtrd(2,0); RRECT(-18,18,18,-18,12,12); RRECT(-16,16,16,-16,10,10);endMXtrd;edge := LNewObj;

result := SubtractSolid(body, hole, difference);result := SubtractSolid(difference, edge, difference);

chrono := GetTickCount-chrono;message(chrono,' /60 equivalent to ', chrono/60, 'seconds');end;

run (test); -----------------

test results on my computer are:with VW8.5: 6/60 equivalent to 0.1 secondswith VW 9 : 970/60 equivalent to 16.16 seconds!!!

The test has been performed on a PowerMac G4 533 MHz, 384Mb phisical RAM, 576Mb virtual RAM122.880 Kb memory assigned to VW8.5160.000 Kb memory assigned to VW9

Both programs have 3D conversion resolution set to minimal in the general preference panel. (Obviously higher 3D resolution means test time increasing).

I'll be glad to know why in VW9 this simple operation is incredibly slower.Please, let me know if the test has similar results on your machines and help me to find out a way out. Thanks Paolo Marcuccetti

--

ARCO ARREDAMENTI srl55049 VIAREGGIO (LU) - ITALY - Via Aurelia Nord, 102tel. +39 0584 46465 - fax +39 0584 46466internet: http://www.arcoarredamenti.it - e-mail: info@arcoarredamenti.it