It seems theres is something wrong in calculating dielectric function in parallel. I tried a few cases like Copper and Silicon. After finishing the groundstate calculation(which is in MPI parallel), the dielectric calculation is performed in serial mode. But the problem is that it freezes out for very long time( for copper case it hanged up calculating for 9 hours! ). I decided to end it up.
But when I do calculation in serial mode for dielectric and momentum matrix, it finishes clean.
Is this a routine procedure or Something I should know better!