eyal@echo.canberra.edu.au (Eyal Lebedinsky) (01/13/91)
Hello again, After anumber of people suggested there is no problem with fp I sat down and did the good old debugging dance. Turns out some prototypes are missing (one was for atof()) and vectors had zero length so the vector normalization divided by zero - simple once you see it. So all is fine. The pc532 runs qrt about 15 times faster than my 20Mhs/286, partly because it has a fp-cop and partly because it IS a fater machine, Now I can get nice pictures in 6 minutes (half size) while in the past I had to use stamp-size for reasonable response. -- Regards Eyal