van@ariel.unm.edu (Van Rauch) (08/09/88)
Has anyone ported spice2g6 to Balance or Symmetry FORTRAN V3.2? Upon making the obvious porting considerations and running it, spice gives: *ERROR*: SYSTEM ERROR, ADDRESS 1677..... IS NOT ON 4 BYTE BOUNDARY One would think that their FORTRAN compiler would know how to align data structures, no? I'm really surprised that more people aren't screaming about lack of runnable third party software for this system. Is it that Sequents' compilers are catching f77 inconsistencies or that Sequent FORTRAN is too radically different? Anxious to get any feedback on this. Also incase you're interested... the function random in Sequents port of /usr/ucb/pc does not produce a uniform distribution of random numbers. Running a chi-square probability test on 52000 tries consistently gives 0.0 probability that the distribution is truly random. The source for random is identical to BSD, save for comments, yet their implementation/assembler/gremlins don't let this thing work true to form. If anyone wants the code to verify this, please let me know. I have reported this to Sequent, (mailbugged it numerous times) and have never received a response. (They have fixed bugs with even numbered case labels and intersection of SETS in pc.) If there are any spice types out there, I'd appreciate any feedback. Van Rauch Staff Consultant University of NM CIRT inet: van@ariel.unm.edu uucp: ...lanl!unm-la!unmvax!ariel!van bitnet: van@unmb