JACOBI:
  CUDA/C version.
  Apply Jacobi's iterative method to solve A*x=b
  System size is N = 10000
  Blocks = 313, threads = 32, B*T=10016

  RMS error ||A*x-b|| = 0.0328983
  RMS error ||x(exact) - x(computed)|| = 0.471175
  Computation required 1.02 seconds

     I    B[i]    X[i]    Xexact[i]  (Xexact-X)[i]  (A*X-B)[I]

     0    1.2860    0.8393    0.8402  9.26e-04  -8.92e-05
     1   -0.8345    0.3926    0.3944  1.76e-03  1.78e-04
     2    0.3734    0.7803    0.7831  2.78e-03  -2.68e-04
     3   -0.0979    0.7949    0.7984  3.52e-03  3.57e-04
     4    0.8273    0.9070    0.9116  4.63e-03  -4.46e-04
     5   -0.8518    0.1923    0.1976  5.29e-03  5.35e-04
     6   -0.2953    0.3287    0.3352  6.48e-03  -6.24e-04
     7    0.9235    0.7612    0.7682  7.05e-03  7.13e-04
     8   -0.7667    0.2694    0.2778  8.33e-03  -8.03e-04
     9    0.3528    0.5452    0.5540  8.81e-03  8.92e-04

JACOBI:
  Normal end of execution.