Can you advise some articles on using fault control in Modelica models ?
For example stop simulation on overheat, electric overvoltage, overload
or (more complex) tracking current parameters in bounds of used model applicability
(for example check signal bounds in case of using low-signal models)

For example I use some basic nozzle model with some parameter R with typical value 0.1,
but extends it with new model of same nozzle with new value R=0.05

I found some strange case in Fritz tutorial, page 65:

slide: Inheriting definitions

record ColorData

    parameter Real red = 0.2;

    parameter Real blue = 0.6;

    Real green;

end ColorData;



class ErrorColor

    extends ColorData;

    parameter Real blue = 0.6;

    parameter Real red = 0.3;   // this line strangly marked as error due to _value_ redefenition

equation

    red + blue + green = 1;

end ErrorColor;

Isn't inheritance must garantee this case as inherited class behavior extension ?

Yes, that's what I had in mind -- dynamic compilation into memory using LLVM JIT low-level objects (vectors, non-C assembly-like functions, command basic blocks), and using some LLVM optimizer passes, triggered by Modelica objects modifications (user modify model -> Modelica engine triggers only this object recompilation in memory in background thread)

Is anybody works on integrating LLVM JIT/dynamic compilation features in OpenModelica ?

Using LLVM in background can avoid large time of model compiling via gcc: every module can be
compiled separately and in background with user continous works. Am I right ?

Or in any case final linkage stage must be done in RAM packing object code for models to work together ?

Is dynamic compilation (triggered by object modifucation) applicable at all for Modelica language realisation ?

.

Does anybody can advise some articles on using OpenModelica in software-domain simulation (maybe with some external tools) ?

I'm interested in embedded software runtime analysis in simulated environment (with modelled peripherial blocks, actuators, async events and realtime-specific software/hardware faults), with AVR8, Cortex-M3 and tiny FPGA/CPLD hardlogic.

I'm first concerned by -msse2 flag mainly, locking down gcc sse codegen to limited sse command set,
but now I think that this selection was done to circumvent some low-level incompatibility in some library or solver, like dassl/x387 said before.

What about SSE3/4 on modern CPU (Intel i7) (gcc -Ofast -march=native -mfpmath=sse ) ?

gcc    -falign-functions -msse2 -mfpmath=sse

Why not used something like gcc -march=native -O3 autodetecting run host optmimizations like -msse avtomatically ?

Does some remote simulation method exists in OpenModelica ?

I have some Window-only workstations and supercluster http://hpc.ssau.ru/
with RedHat Linux / x5560 nodes powered with some MPI-enabled libraries.

Where can I get info on remote compiling & simulation OM models
with some details on configuring and SSH-connection ?

Can't find samples on using texcell - direct input like \LaTeX causes error:

! LaTeX Error: Missing \begin{document}.

Adding \documentclass{article}\begin{document}\LaTeX\end{document}

gives result with very large mergins and toital image size.

How I must wrap my texts and furmulas to properly format texcells ?