Directed floating point rounding?

[Thomas Epperly]

How do you change the IEEE floating point flags on the DECstation 3100 in C?
I want to change the rounding mode for a numerical application.  I found
some clues in /usr/include/fpu.h and some routine definitions, but they
don't have any explanations.  I think that knowing the parameters to
get_fpc_csr() and set_fpc_csr would be all that I need.  Please send any
information or suggestions.

Tom Epperly
te07 at cs.cmu.edu
epperly at osnome.che.wisc.edu
--