Large system simulation problem

[Alaba]

Hello there,

I'm building a system that includes a large number of resistors.  I used  sc_vector  of elements and nodes, as I increase the number of resistors I get segmentation error.

The below example is a simpler form  of my code that also gives the same error  

 

In the code below if I use small value of N:

( N=10000):  I get the right result 

but for N=1048576:  //(large N)

I get: Segmentation fault (core dumped) 

 

------------------------------------------------------resistors in series module----------------

//            p-    r0-r1-r2-r3...rN   -n

 static const int N=1048576;
  
    sca_eln::sca_terminal n;
    sca_eln::sca_terminal p;
   
    sc_vector< sca_eln::sca_node > c_vec{"c_vec", N };  //nodes beteen resistors  
    sc_vector<sca_eln::sca_r>     rs_vec{"rs_vec",   N  };  

SC_CTOR(rmatnn):  
                            
       p("p"),n("n") // 
                
      {

    rs_vec[0].p(p); //connect  p  port of first resistor to main p port 
    rs_vec[0].n(c_vec[0]); //connect n port of first resistor to nod 0
    rs_vec[0 ].value=10;
    
    for (int i = 1; i < N-1;  i++) {
      rs_vec[i ].p(c_vec[i-1]);
      rs_vec[i ].n(c_vec[i]);
      rs_vec[i ].value=10;
          }
     rs_vec[N-1 ].p(c_vec[N-2]);
     rs_vec[N-1 ].n(n);
         rs_vec[N-1 ].value=10;
        } 
      };

Is there any limitations on the number of modules we use  for the simulation??

Thanks in advance 

 

[karsten]

There is no real limitation except the available memory. Each resistor is a SystemC module which allocates some memory due the sc_module members.

I guess SystemC and SystemC AMS does not check always, that memory could be allocated - thus you get the segfault.

You can try a computer with more memory or try to setup the equation manually- for a resistive network this should be some lines of c code only - if you have inductors or capacitors you can use the state space or Ltf objects.

[Alaba]

12 hours ago, karsten said:

There is no real limitation except the available memory. Each resistor is a SystemC module which allocates some memory due the sc_module members.

I guess SystemC and SystemC AMS does not check always, that memory could be allocated - thus you get the segfault.

You can try a computer with more memory or try to setup the equation manually- for a resistive network this should be some lines of c code only - if you have inductors or capacitors you can use the state space or Ltf objects.

I get same problem for same value of N with system of 32GB RAM and 8GB Ram 

[karsten]

I just saw, that it looks like you instantiate the modules statically. You may run into the OS stack size limitation. Allocate the modules/vectors dynamically (using new) may solves the problem (or at least increases the possible size).

May you can try similiar like this (not checked):

    sc_vector< sca_eln::sca_node >* c_vec;  //nodes beteen resistors  
    sc_vector<sca_eln::sca_r>*     rs_vec;  

SC_CTOR(rmatnn):  
                            
       p("p"),n("n") // 
                
      {

c_vec=new sc_vector< sca_eln::sca_node >("c_vec", N );

rs_vec=new sc_vector<sca_eln::sca_r>("rs_vec",   N  );

 ...
      };

 

what are the shell messages before the segfault?