I have ported some code from Sparc to Intel. The code works fine except it prints NaNs for some floats that are not NaNs at all. One of my co-workers suggested I run the same code (of course compiled) on FreeBsd and see if the NaNs appear. The FreeBsd code does not produce the NaNs. I have searched the Web for Gcc and Nans and the closest I have come accross was a mention of this problem on the Alpha chips and not Intel processors. (Does that mean Intel copied that much from Digital that they too have NaNs for usual float? :) :) )
So any expert out there where should I look further? /----------------------------------------------------------------------------/ Daniel J. Mashao -- [EMAIL PROTECTED] http://www.ee.uct.ac.za/~daniel /----------------------------------------------------------------------------/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
