David LaFrance-Linden wrote:
>Small nit. When I brought up canonical values and the testing
>methods, I never said -1. I said "all ones." It is more proper to
>say (in C) ~0 (and I can't find the unary element function for bit
>complement in F90). This works for both two's complement and one's
>
Bitwise complement is NOT(i); see section 13.7.87 in the standard.
Cheers,
Bill
>complement machines, and is more correctly aligned with the intent
>(assuming 0 is represented by all zeros).
>
>
>
--
Bill Long [log in to unmask]
Fortran Technical Support & voice: 651-605-9024
Bioinformatics Software Development fax: 651-605-9142
Cray Inc., 1340 Mendota Heights Rd., Mendota Heights, MN, 55120
|