pentium bug found by Kahan's number-theory methods if you wait long enough
uunet!cwi.nl!Dik.Winter
uunet!cwi.nl!Dik.Winter
Thu Nov 17 15:29:52 PST 1994
Interestingly enough just today there was an article in the comp.arch
newsgroup (and probably others) that had a (fairly small) program that
mimics Pentium division, including the bug.
dik
More information about the Numeric-interest
mailing list