ALG106@psuvm.psu.edu (02/26/91)
My brother is writing a program for his company that requires floating
point operations. The program works flawlessly when running under
the emulation library (w/o coprocessor) but crashes when running with
a coprocessor. The error he gets is an exception 6. If it makes a
difference he wrote the program in Microsoft C 6.??.
I don't know the specifics of the problem, but would greatly appreciate
it if someone who is familiar with the 80387 could call my brother and
perhaps offer some advice to fix this problem. He found the problem
on Saturday and has been working on this problem all day Sunday and Monday.
If you can think of anything that could help, or would like to talk to him,
send me E-Mail and I will send you his home and work phone numbers.
Let me just say again that I do not understand the problem - I've never
programed math coprocessors, or used C that much. So please, if you can
help I will get you in touch with my brother.
Thank you in advance.
Andy
------------------------------------------------------------------------
Andrew Gross BITNET: ALG106@PSUVM
Penn State University INTERNET: alg106@psuvm.psu.edu
alg@eclu.psu.eduebergman@isis.cs.du.edu (Eric Bergman-Terrell) (03/01/91)
I developed an application that does a great deal of floating point - it uses software floating point or the math chip if present. The biggest problem that I had - my software had many "large" floating point expressions - some of which went on for pages. I had to simplify some of them because they overflowed the math coprocessor's operand stack. If you can figure the line where the program is crashing, try to simplify the expression and see if that works. Terrell