FDS Update

["Kevin McGrattan" <kevin.mcgrattan@nist.gov>]

All

I haven't written for a while, so I thought it would be a good idea to 
update you on progress towards a new version of FDS/Smokeview.  We had hoped 
to have a new version by the end of the year, but that is not going to 
happen.  The delay is mainly related to FDS, not Smokeview.  As FDS matures, 
there just seems to be more to do to get out a new release.  Certainly there 
are improvements to the various solid and gas phase routines, but there's 
also the issue of verification, validation, and quality control.  We don't 
want the new version of FDS to be buggier than the old, even though we've 
had to change much of the code to accommodate new features.  So be patient.

The good news is that we have hired a new staff member at NIST, Bryan Klein, 
for much needed support.  Bryan worked at the Western Fire Center for 5 
years, where he did everything from data acquisition to FDS modeling.  He 
will provide for us the perspective of the user, as opposed to the 
developer.  His first task is to overhaul the User's Guide and everything 
associated with running the model.  His specialty is IT, so he will drag us 
into the 21st Century by introducing us to modern code development 
techniques, archiving, bug reporting, and so on.  Up to now, we've developed 
FDS in a fairly ad hoc manner, but we want to take advantage of various 
tools that are available for streamlined code development and maintenance. 
And even with the various FDS interfaces, whether commercial or private, 
there are still many improvements to be made to make FDS easier to use, and 
to make FDS "interface" with the interfaces more cleanly.  We're also 
struggling just to keep up with the latest technology.  Dual cores, Quad 
cores, Opterons, Xeons, MPI, and goddammit why won't it run on my Mac!  That 
sort of stuff.

So we'd like to welcome Bryan to the job by asking anyone with a suggestion 
for the new FDS to send email to bryan.klein@nist.gov with the subject line 
"FDS Suggestion".  Some of you have already done this, so there is no need 
to resend.  Suggestions for Smokeview should still be addressed to 
glenn.forney@nist.gov.

Jason Floyd of Hughes Associates and I are working on improvements to the 
gas phase combustion routines, so any comments or suggestions in this regard 
can be sent to me or jfloyd@haifire.com.  Simo Hostikka of VTT, Finland has 
been working on the solid phase routines.  He can be reached at 
simo.hostikka@vtt.fi.

At some point as we near release in early 2007, I will ask that those of you 
with persistent, irritating bugs to prepare simple input files that 
demonstrate the bug, so that we can check to see if we've either fixed it, 
or if it is something that still needs work.  There will always be scenarios 
or phenomena that we can't address, but we want to fix problems are simply 
due to programming bugs.  Keep in mind that the phrase "simple input file" 
literally means 10 to 20 lines that clearly demonstrate a problem.  No doubt 
that there are calculations that do funny things after 2 weeks running on 16 
processors, but it's hard to debug something like that.  It's much better to 
have a short, quick calculation that demonstrates the same problem.

Kevin


Kevin McGrattan
National Institute of Standards and Technology
100 Bureau Drive, Mail Stop 8663
Gaithersburg MD 20899
Phone: 301 975 2712
Fax:       301 975 4052