g77 bug, double precision

Dave Love d.love@dl.ac.uk
Thu Jan 7 12:43:00 GMT 1999


>>>>> "ACR" == Arturo C Ramos <arturo_c_ramos@amoco.com> writes:

 ACR> The following definition
 ACR> 	real *8 x /1E15/, y /1D15/

 ACR> would not be equal after compiling.

 ACR> x would have a value of 9.99999987E+14
 ACR> while y would have the correct value of 1.0E+15

See the node `Floating-point errors' in the g77 manual concerning
floating point equalities.



More information about the Gcc-bugs mailing list