I know many of you are wondering what has happened to the next release of
the LCD tools. We have been suffering a lot of operating system/compiler
versionitis problems which have slowed us down. The current state of play
is that we have constructed what we believe to be a fully corrected set of
Gismo/GismoApps tools (but still with last year's detector configurations).
We are in the process of testing them:
Platform Comment
AIX Probably fully functional. We suffer the occasional lockup of
jobs, but this is probably due to the machine configuration
(available swap space) rather than the simulation code.
SunOS Fully functional.
Linux Fully functional WHEN COMPILED NON-OPTIMIZED. There is a bug in
the egcs/gcc optimizer for x86 platforms. There is believed to
be a patch for this which we are investigating.
DEC/OSF1 Bombs out on event 6 of our test dataset (100 t-tbar events).
Cause not known. May be machine configuration (again). We
only have access to a small and underpowered DEC/OSF1 box.
Windows/NT Started working this morning. Runs to completion.
We have compared the simulated datasets (crudely) from AIX, SunOS and Linux.
They appear to be the same. We are refining the test program and will add
the Windows/NT dataset to the comparison.
So our priorities are:
1) Fix whatever is wrong with DEC/OSF1 (it would be DEC/OSF1 of course ...
we realize that our DEC/OSF1 customers are particularly impatient).
2) Try to build a working, optimizable Linux environment (for those into
the technicalities we're targetting RedHat 6.0 (with the new 2.2 kernel)
and gcc 2.95.1 with the optimization patch).
Now if our computer centre could just stop having power outages...
Thank-you for your patience.
Anthony Waite.
