i91@nikhefh.uucp (Fons Rademakers) (05/09/86)
We of the high energy physics community favor the Apollo because
of its extremely nice and fast FORTRAN compiler. I don't have experience
with the Sun, but if it is full bsd4.2 it also means it has the lousy
f77 compiler, which is a genuine horror compared to Apollo's ftn.
Apollo is also doing well with its new DN580 in the field of high
speed graphics (animation) machines. The performance of the DN580 comes
close to that of the fast Megateks, which are currently used
for scanning of the physics events obtained from the large
detectors and accelerators in CERN, Geneva. This means that the DN580
(which we already benchmarked) might be a good replacement of the Megateks.
By the way SR9.2 solves a lot of Unix problems of SR9.0. For example
the ps command is now as fast as on a "real" Unix system.
Greetings,
--
Fons Rademakers.
NIKHEF-H, National Institute for Nuclear and High-Energy Physics.
UUCP: {decvax,cernvax,okstate,garfield,philabs,seismo}!mcvax!nikhefh!i91
p-mail: Nikhef-H \ /
Kruislaan 409 \__________/
P.O. Box 41882 / \
1009 DB Amsterdam - Netherlands / \