About building FMU-QSS
Closed this issue · 7 comments
The FMU-QSS support is not currently in use by our project so it wasn't up to date. I updated it to match the current code base so please give it a try now.
Unless you want to use QSS on multiple, connected FMU-ME FMUs the FMU-QSS support won't be useful.
OK, thanks
Hello, now, i have a modelica model , how can i simulate my model with this solver ? please give me some ideas.
another issue :i pull your latest version, when building, it failed.
errors like :
/home/cbn/Desktop/project/SOEP-QSS/src/QSS/Observers.hh:1075:62: error: dependent-name
/‘QSS::Observers::Variables::size_type’ is parsed as a non-type, but instantiation yields a type
1075 | for ( Variables::size_type i = 0, e = qss_observees_.size(); i < e; ++i ) { // Set observee directional derivative vector
| ^
/home/cbn/Desktop/project/SOEP-QSS/src/QSS/Observers.hh:1075:62: note: say ‘typename QSS::Observers::Variables::size_type’ if a type is meant
Hello, now, i have a modelica model , how can i simulate my model with this solver ? please give me some ideas.
The current SOEP-QSS is designed to work with FMUs built from Modelica models using Modelon's latest Optimica Compiler Toolkit that has some options/features for QSS. (Earlier revisions of this repository could work, with limitations, with JModelica or Dymola.)
This information is in the README.md file.
another issue :i pull your latest version, when building, it failed. errors like :
/home/cbn/Desktop/project/SOEP-QSS/src/QSS/Observers.hh:1075:62: error: dependent-name /‘QSS::Observers::Variables::size_type’ is parsed as a non-type, but instantiation yields a type 1075 | for ( Variables::size_type i = 0, e = qss_observees_.size(); i < e; ++i ) { // Set observee directional derivative vector | ^ /home/cbn/Desktop/project/SOEP-QSS/src/QSS/Observers.hh:1075:62: note: say ‘typename QSS::Observers::Variables::size_type’ if a type is meant
This is corrected in the current master branch. The Visual C++ and Intel C++ compilers on Windows (unlike GCC) are lax about places like this where typename
is technically required.
ok, thanks a lot !