[comp.os.vms] Fortran-77 VAX to RSX

EVERHART%ARISIA@rca.COM.UUCP (08/26/87)

The DEC F77 compiler for RSX can readily be built under VAX-11 RSX
and runs just fine there. It'll run on ANY VAX that runs VAX-11 RSX.
BTW you can just get the F77 compiler kit for RSX and use that; only
difference should be packaging. It's pretty easy to figure out.
	A suggestion: When you get it, EDIT the F77 taskbuild file
before running it. You may want to change some defaults or increase
some default limits. You'll NOT want I/D space support, for instance,
under VAX-11 RSX; no point in having it. I also recommend using the 
FCS version of the Fortran library rather than the RMS version for
most things. You can have both libraries, but the FCS one takes less
address space, which is important in compat mode. Both run just fine on
any VAX. VAX-11 RSX also works well, and has been improving lately,
in handling most RSX user mode code. Lots of things from DECUS that
were designed all or part in macro-11 run just fine under it, and
many do jobs that one needs occasionally done. I keep the code around
in that form because it runs, and for infrequent use it isn't worth
converting to native mode. Remember too, that I/O is always done at
bottom in native mode; I/O intensive applications with a little
compat mode computing are not necessarily CPU hogs. On the other
hand, your Monte Carlo calculations should not be run in compat mode
unless you have NO other choice.
	I've moved a fair number of applications from RSX to VAX by
recompiling and some the other way. FYI, be aware that the RSX SIG tapes
of DECUS have lots of interesting PDP11 code in HOLs which will either
compile native mode directly, or will compile and run with one or
two assignments or edits (typically you may want to change TI: to
TT: or assign TT: to TI: so terminal I/O will run unaltered). Much
of this applications code appears NOWHERE else in DECUS and it's a
shame so many VAX sites ignore it because of its' PDP11 origins. I'll
invite the rest of you folks to look over the abstracts in the DECUS
library catalog. After reading, you'll be able to identify a good
bit of this. You'll find that EVERY recent RSX SIG tape has a significant
amount of such code. The Spring '87 tape has even more than usual.
'Nuff said...

Glenn Everhart
Everhart%Arisia.decnet@ge-crd.arpa