rab@cdcvax.UUCP (Roger Bielefeld) (05/02/84)
Recently I have been involved in porting an extremely large FORTRAN program to f77 on a 4.1BSD 11/750. Performance of the final product is an issue: execution speed of the UNIX version must be within 10 percent of the VMS version, except that slowness of execution that can be attributed to the f77 compiler doesn't count. Now, naturally, I'll have to demonstrate that the f77 compiler is at fault when the final timings are done. So, finally, to my question: Does anybody have any information comparing the performance of the two compilers in question? How about the nature of the differences in execution speeds? Are they caused by inefficient references to subroutine arguments, to inefficiencies in subroutine calls, or what? Please respond to the path below, not the path in the header. If anybody else is interested in any information I get, drop me a line and I'll be happy to forward you a copy. Thanks. -- Roger Bielefeld decvax!cwruecmp!rab (UUCP) rab.case@csnet-relay (ARPA) rab@case (CSNET)