Diagonally dominent matrix with desired condition number

I have parallel code for Jacobi algorithm and I want to test it on a square large linear system(like 1600*1600). I tried with symmetric positive definite but the solution does not converge. Could help with generating a diagonally dominant matrix with desired condition number(like 1, 1000 , 10000 , 100000). I tried with this code but it does not work:

function result = matrix_jacobi(n , CondNumb )

tempA  = round(1000*randn( n, n));

A=  tempA  + diag(sum(abs(n ),2));

[U,S,V ] = svd(A);
S(S~=0)=linspace(CondNumb , 1 ,  n);
result =U*S*V';