From: Gary Funck (gary_at_intrepid_dot_com)
Date: Fri Jul 17 2009 - 11:17:35 PDT
On 07/17/09 16:56:21, sainath l wrote: > Thank you for the reply. > Actually I am not worried about the performance here. I am interested in > finding out how much time the allocation routine in itself takes. > > I would need an array of pointers like > > shared int *shared a[N]; As others have pointed out, it isn't clear that a shared pointer to an array of pointers to rows is the best way to express a shared matrix in UPC, because it may lead to poor performance, and leaves open the question of thread affiinity. I've attached a simple program that illustrates dynamic allocation of an NxM array using the method of representing 'a' that you've suggested above. Note that all allocated elements have affinity to thread 0 in this implementation, and that is likely not a good representation for a real world problem. - Gary