Christoph Pospiech <christoph.pospi...@de.ibm.com> wrote on 09/07/2010
03:18:51 PM:
> Hi,
>
> I was trying to do the equivalent of an MPI_Gather using code like
this.
>
> finish for ( pt(i) in v_dst ) {
> v_loc(i).copyFrom(0,v_dst(pt),0,vsize);
>
> where
> static type Array1D = Array[Double]{rank==1};
>
> global val v_dst: DistArray[Array1D]{rank==1};
> global val v_loc: Array[Array1D]{rank==1};
>
> - the region for all objects of type Array1D is [0..vsize-1],
> - the region for v_loc and v_dst is [0..number_of_places-1],
> - the distribution for v_dst maps exactly one point to each place.
>
> The above code runs fine on one place. But on two places, it throws a
> x10.lang.BadPlaceException: point (1) not defined at (Place 0).
> pt(1) lives on place 1 by definition, so the exception is thrown
correctly.
> Means - the above code snippet is wrong. I tried about half a dozen
variants
> of the above, but no avail.
>
> How to do it right ?
In 2.0.6 you can do
finish ateach (pt(i) in v_dst) {
val t = v_dst(pt); // note that v_dst(i) is slightly more
efficient
at (v_loc) v_loc(i).copyFrom(t);
// or, alternatively (but not as efficiently)
// v_dst(pt).copyTo(at (v_loc) v_loc(i));
}
(the above copies the whole array).
Igor
--
Igor Peshansky (note the spelling change!)
IBM T.J. Watson Research Center
X10: Parallel Productivity and Performance (http://x10-lang.org/)
XJ: No More Pain for XML's Gain (http://www.research.ibm.com/xj/)
"I hear and I forget. I see and I remember. I do and I understand" --
Confucius
------------------------------------------------------------------------------
This SF.net Dev2Dev email is sponsored by:
Show off your parallel programming skills.
Enter the Intel(R) Threading Challenge 2010.
http://p.sf.net/sfu/intel-thread-sfd
_______________________________________________
X10-users mailing list
X10-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/x10-users
