FDS6 run using FDS 5.5.3
536 views
Skip to first unread message
Erichovi
Jul 6, 2012, 2:32:32 AM7/6/12
to fds...@googlegroups.com
The other day I started a couple of simulations with FDS6=.TRUE. on the misc line to compare the results I got when running on FDS 5.5.3 SVN 7031 (for WIN7 64bit). However, in the .out files for the FDS6 simulations, I get the following:
Compilation Date : Fri, 29 Oct 2010
Version : 5.5.3 Parallel
OpenMP Disabled
SVN Revision No. : 7031
So... does this mean I actually got the results for FDS 5.5.3 or was FDS6 actually activated and used in the simulation? I understand that the .out files would not have a different version of FDS than I have installed (being 5.5.3) but what confuses me is that the validation files on the google code fds-smv website (eg. http://code.google.com/searchframe#xbii4fg5bFw/trunk/FDS/trunk/Validation/Vettori_Sloped_Ceiling/FDS_Output_Files/Vettori_Sloped_13SFC.out) has:
Compilation Date : Fri, 02 Sep 2011 Version : 6.0.0 Serial OpenMP Disabled SVN Revision No. : 8990
I always thought FDS6, though still unreleased officially, was included in the 5.5.3 download from the NIST website?
Thanks for your help!
Erik
dr_jfloyd
Jul 6, 2012, 8:02:35 AM7/6/12
to fds...@googlegroups.com
At the time of FDS5.5.3, the FDS6 flag was nothing more than a way for the development team to be able to access features being developed for FDS6. The flag invoked a parallel execution path through some of the subroutines. The FDS6 flag is no longer in the source code as we no longer plan another FDS5 release and are focusing efforts on FDS6. Much of what the FDS6 flag activated in FDS5.5.3 is either no longer present in the current code or has changed greatly. The current trunk on GoogleCode represents the current state of the project which is the source that will be released as FDS6. So all of the V&V files have been updated to reflect that. The V&V files for FDS5.5.3 can still be obtained by downloading SVN 7031.
Michael K
Jul 6, 2012, 9:49:39 AM7/6/12
to FDS and Smokeview Discussions
Dr. Floyd,
I am not certain that I understand this response. The latest release
of FDS is version 5.5.3 from Oct. 2010. That is the version that
practitioners in industry are still using and will continue to use
until NIST decides to do another release. Does the FDS6 flag in that
version call in different models than would normally be called in a
simulation where FDS6 is set to .FALSE.?
We have already discovered that setting the DYNSMAG flag to .TRUE. can
cause an error in the calculation of surface heat release rate and
while that error is known to the developers, a correction has not been
released. If the FDS6 flag also invokes the dynamic Smagorinsky LES
model then it could cause the same error. It seems that the FDS6 flag
either does not have a function in release 5.5.3 or if it does it can
introduce error into calculations and should not be used.
Would you clarify this please?
Thank you,
Michael
On Jul 6, 7:02 am, dr_jfloyd <drjfl...@gmail.com> wrote:
> At the time of FDS5.5.3, the FDS6 flag was nothing more than a way for the
> development team to be able to access features being developed for FDS6.
> The flag invoked a parallel execution path through some of the
> subroutines. The FDS6 flag is no longer in the source code as we no longer
> plan another FDS5 release and are focusing efforts on FDS6. Much of what
> the FDS6 flag activated in FDS5.5.3 is either no longer present in the
> current code or has changed greatly. The current trunk on GoogleCode
> represents the current state of the project which is the source that will
> be released as FDS6. So all of the V&V files have been updated to reflect
> that. The V&V files for FDS5.5.3 can still be obtained by downloading SVN
> 7031.
>
>
>
>
>
>
>
> On Friday, July 6, 2012 2:32:32 AM UTC-4, Erichovi wrote:
>
> > The other day I started a couple of simulations with *FDS6=.TRUE.* on the
> > * Version : 5.5.3 Parallel*
> > *
> > *
> > * OpenMP Disabled*
> > *
> > *
> > * SVN Revision No. : 7031*
> > has:
>
> > * Compilation Date : Fri, 02 Sep 2011** Version : 6.0.0 Serial**** OpenMP Disabled**** SVN Revision No. : 8990**
> > *I always thought FDS6, though still unreleased officially, was included in the 5.5.3 download from the NIST website?
I am not certain that I understand this response. The latest release
of FDS is version 5.5.3 from Oct. 2010. That is the version that
practitioners in industry are still using and will continue to use
until NIST decides to do another release. Does the FDS6 flag in that
version call in different models than would normally be called in a
simulation where FDS6 is set to .FALSE.?
We have already discovered that setting the DYNSMAG flag to .TRUE. can
cause an error in the calculation of surface heat release rate and
while that error is known to the developers, a correction has not been
released. If the FDS6 flag also invokes the dynamic Smagorinsky LES
model then it could cause the same error. It seems that the FDS6 flag
either does not have a function in release 5.5.3 or if it does it can
introduce error into calculations and should not be used.
Would you clarify this please?
Thank you,
Michael
On Jul 6, 7:02 am, dr_jfloyd <drjfl...@gmail.com> wrote:
> At the time of FDS5.5.3, the FDS6 flag was nothing more than a way for the
> development team to be able to access features being developed for FDS6.
> The flag invoked a parallel execution path through some of the
> subroutines. The FDS6 flag is no longer in the source code as we no longer
> plan another FDS5 release and are focusing efforts on FDS6. Much of what
> the FDS6 flag activated in FDS5.5.3 is either no longer present in the
> current code or has changed greatly. The current trunk on GoogleCode
> represents the current state of the project which is the source that will
> be released as FDS6. So all of the V&V files have been updated to reflect
> that. The V&V files for FDS5.5.3 can still be obtained by downloading SVN
> 7031.
>
>
>
>
>
>
>
> On Friday, July 6, 2012 2:32:32 AM UTC-4, Erichovi wrote:
>
> > misc line to compare the results I got when running on FDS 5.5.3 SVN 7031
> > (for WIN7 64bit). However, in the .out files for the FDS6 simulations, I
> > get the following:
> > * Compilation Date : Fri, 29 Oct 2010*
> > (for WIN7 64bit). However, in the .out files for the FDS6 simulations, I
> > get the following:
> > * Version : 5.5.3 Parallel*
> > *
> > *
> > * OpenMP Disabled*
> > *
> > *
> > * SVN Revision No. : 7031*
>
> > So... does this mean I actually got the results for FDS 5.5.3 or was FDS6
> > actually activated and used in the simulation? I understand that the .out
> > files would not have a different version of FDS than I have installed
> > (being 5.5.3) but what confuses me is that the validation files on the
> > google code fds-smv website (eg.
> >http://code.google.com/searchframe#xbii4fg5bFw/trunk/FDS/trunk/Valida...)
> > So... does this mean I actually got the results for FDS 5.5.3 or was FDS6
> > actually activated and used in the simulation? I understand that the .out
> > files would not have a different version of FDS than I have installed
> > (being 5.5.3) but what confuses me is that the validation files on the
> > google code fds-smv website (eg.
> > has:
>
> > * Compilation Date : Fri, 02 Sep 2011** Version : 6.0.0 Serial**** OpenMP Disabled**** SVN Revision No. : 8990**
> > *I always thought FDS6, though still unreleased officially, was included in the 5.5.3 download from the NIST website?
dr_jfloyd
Jul 6, 2012, 10:01:49 AM7/6/12
to fds...@googlegroups.com
FDS 5.5.3 is the latest official release of FDS. The FDS6 flag was added by the development team to allow us to test new features without breaking the release version, so yes enabling FDS6 turned on experimental features that were different than the official 5.5.3. DYNSMAG, FDS6, and a number of other undocumented commands in 5.5.3 represent capabilities under development. That FDS6 or DYNSMAG may not result in reasonable simulations is not surprising given that they are experimental.
Michael K
Jul 6, 2012, 5:17:55 PM7/6/12
to FDS and Smokeview Discussions
Dr. Floyd,
Thank you.
Thank you.
Erichovi
Jul 8, 2012, 7:47:15 PM7/8/12
to fds...@googlegroups.com
Thank you for the help Jason, that all makes sense. Is there any way for us non-developers to test run FDS6 now? I was reading in an old thread (https://groups.google.com/d/msg/fds-smv/e7NXGXhHVrc/0JljhLGWsn4J) that "FDS 6 options are available with the latest SVN of the code", however that is obviously not the case anymore. Or at least, the latest FDS6 options are obviously not included in SVN 7031.
dr_jfloyd
Jul 8, 2012, 8:04:44 PM7/8/12
to fds...@googlegroups.com
FDS is open source. You are free at any time to download and compile the latest source.
Michael K
Jul 9, 2012, 9:42:28 AM7/9/12
to FDS and Smokeview Discussions
Dr. Floyd,
For those of us who are not in academia, that really is not an
option. It certainly is not free as there is the cost of the compiler
and the opportunity cost of not serving clients with that time used to
create and validate our own nonconfigure-controlled executables that
must be accounted. When is the anticipated release date for FDS6?
Thanks,
Michael
For those of us who are not in academia, that really is not an
option. It certainly is not free as there is the cost of the compiler
and the opportunity cost of not serving clients with that time used to
create and validate our own nonconfigure-controlled executables that
must be accounted. When is the anticipated release date for FDS6?
Thanks,
Michael
dr_jfloyd
Jul 9, 2012, 10:21:05 AM7/9/12
to fds...@googlegroups.com
The question was how could one test run FDS 6. Whether or not it is convenient or practical does not change the answer that if you wish to work with the latest developmental version, you need to download the source and compile it.
We do not have a target date for the release of FDS 6. Please read: http://blog.fds-smv.net/2011/02/progress-on-fds-6.html
While that blog post is a little over a year old, it still reflects where we are with FDS 6. In addition to new capabilities, many changes have been made under the hood with an eye towards making FDS easier to maintain and more flexible for future development. The process of testing and debugging those changes, developing V&V for the changes, doing documentation, plus fixing other issues that have arisen in the process takes a considerable amount of time. Keep in mind that we are not Ansys or CD-Adapco with a multitude of software developers on staff. We have made a lot of progress, and I am hopeful for at least a beta release sometime later this year (and hopefully Kevin and Randy concur with that).
We do not have a target date for the release of FDS 6. Please read: http://blog.fds-smv.net/2011/02/progress-on-fds-6.html
While that blog post is a little over a year old, it still reflects where we are with FDS 6. In addition to new capabilities, many changes have been made under the hood with an eye towards making FDS easier to maintain and more flexible for future development. The process of testing and debugging those changes, developing V&V for the changes, doing documentation, plus fixing other issues that have arisen in the process takes a considerable amount of time. Keep in mind that we are not Ansys or CD-Adapco with a multitude of software developers on staff. We have made a lot of progress, and I am hopeful for at least a beta release sometime later this year (and hopefully Kevin and Randy concur with that).
Kevin
Jul 9, 2012, 11:11:34 AM7/9/12
to FDS and Smokeview Discussions
Well said. We're working towards a beta release of FDS 6 this summer.
Reply all
Reply to author
Forward
0 new messages