Re: Can't get exact solution. Why ?



with LAPACK and the g95 compilers here is what I get
X1= 0.21428571428571416
X2= 0.07142857142857154
X3= 0.14285714285714285
X4= 0.14285714285714285
X5= 0.07142857142857154
X6= 0.21428571428571416

Check, may be the sparse implementation is wrong. The matrix is
triadiagonal, so you should be using the triadiagonal
matrix decomposition

.