I changed the For loop for a compile

rho = Compile[{{xp, _Real, 1}, {xg, _Real, 
    1}, {carga, _Real}, {np, _Real}, {ng, _Real}, {dx, _Real}},
  n = ng + 1;
  \[Rho] = Table[carga*np/(dx*ng), n];
  \[Rho][[ng + 1]] = 0;

  Do[If[Abs[xp[[i]] - xg[[j]]] <= 
     dx, \[Rho][[j]] = \[Rho][[j]] - 
      carga*(1 - (Abs[xp[[i]] - xg[[j]]])/dx)/dx; r += 1, 0] , {i, 1, 
    np}, {j, 1, ng + 1}];
  \[Rho][[1]] = \[Rho][[1]] + \[Rho][[ng + 1]];]

Is it possible to optimize even further?

Hi In my code I have 2 For loops like that and will have a sow and reap algorthim (haven't done that part yet) to build a Table where I storage a list (very simillar to [Rho]) in each iteration. And I think I will need a big number of iteratios. So I am trying to make the code as fast as possible. (if you have a fast solution to implement this storage algorithm I am also interested)

While I was experimenting with nparticles=10 000 and ngrid=1000 that For loop took 15 secs. That was unaceptable. With the compile version it takes around 1.5 secs. So if there are any more good solutions to optimize the code I am very interested. As I think these numbers for nparticles and ngrid are as high as I am interested to do and they might be too high I am trying to make a functional code for them.

Thanks for the information about the dist function.