> >> We do have an "engineering exercise" in progress to create a native
> >> DIG..er.. Compaq Fortran compiler and associated libraries on Linux for
> >> Alpha (not X86)systems, and it's going very well so far. The compiler
> >> work with existing Linux development tools (gdb, etc.) - we're not
> > ^
> >Is the "existing" gdb going to support the full Fortran 95 syntax?
> No. We're looking at porting our ladebug de*, but that's more
> speculative at this time. We figure that what people want most is a fast,
> robust Fortran 95 compiler on Alpha Linux and that getting such a thing out
> there soon is more important than delivering a "complete Fortran development
> environment" later.
> DIGITAL Fortran Development http://www.veryComputer.com/
> Compaq Computer Corporation
> 110 Spit Brook Road, ZKO2-3/N30
> Nashua, NH 03062-2698 "Free advice is worth every cent"
> DIGITAL Fortran web site: http://www.veryComputer.com/
DU er.. T64... and DF. In fact, I went through g77 benchmark test
with the Polyhedron F77BENCH. P-II at 400MHz/g77 was factor of
~1.2 faster than Linux/Alpha/g77 with 21164LX-4 at 533MHz. So,
compared to DF for Windows 98 which is 1.5 times faster than g77/P-II,
g77/Alpha21164/Linux is a complete loser. 21264 is at present
out of reach for most of us.
See how much people get e*d with your reply. Compaq
should give your team a good deal of bonus this Christmas. I am not
sure whether your higher managment will know your effort for "saving"
Alpha or not.
BTW, I found that any benchmark program with exp(.) or sqrt(.) makes
Alpha/g77 sweat and puff, because they have not been implemented in
libffm yet. Fotunately, my eigenvalue code still runs faster on Alpha/g77
than on P-II.
G. Hugh Song