I'd like to thank everyone who has recently tested some software of
mine
http://atlc.sourceforge.net/
for computing the properties of arbitrary shaped transmission lines
and couplers. Hopefully this software, which is free for amateur or
commercial use, is of benefit to the amateur radio community.
So thanks to Daniel Schaefer, Bdale Garbee KB0G, Steve Thompson G8GSQ,
Ron Patterson W6FM, David Anderson GM4JJJ and Roger Rehr W3SZ for
testing this very recently. Sorry if I forgot anyone.
A couple of you with SuSe Linux had a problem, which was easily fixed.
I have released an updated version 4.4.0, which has some quite major
changes. These changes were made because the code would not compile on
a rather old Cray Y-MP supercomputer, since the size of all data types
(including short) are 8 bytes on the Cray. This has been fixed, but
since the changes needed to allow it to run on the Cray supercomputer
were quite considerable, other bugs might have crept in.
If anyone who has a UNIX or Linux computer could test it I'd be
grateful. I'm particularly interested if anyone has access to less
common or older hardware/software.
configure --with-threads
make
make check
will do it all. It would be helpful if the output was directed to a
file
configure --with-threads > callsign.log
make >> callsign.log
make check >> callsign.log
and I was emailed callsign.log. If your callsign is on the log file, I
won't loose who sent me the log files. Let me know your hardware,
operating system and compiler.
Note, my previous request to add the option --enable-hardware-info is
absent, as I've now made that the default.
PS
The last test runs a benchmark. It is clear that some of you are
getting quite poor performance from quite modern hardware, suggesting
your choice of compiler flags is not optimal and therefore you might
be compiling ALL your programs less than optimally. The time to run
the benchmark on my 400 MHz HP C3000 PA-RISC processor is 157 s and
it's 86 s on a Sun Ultra 80 workstation with 4x450 MHz CPUs,
increasing to 265 s if only one of the four processors are used.
Bother machines are dicontinued models that are several years old, so
modern Pentiums and Athlons should do a lot better.
PLEASE DONT SEND TEST RESULTS TO THE MAILING LIST - SEND THEM DIRECTLY
TO ME.
--
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Website: http://www.medphys.ucl.ac.uk/~davek
Author of 'atlc' http://atlc.sourceforge.net/
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|