Oops. Josh is correct -- the code won't work as written, even if real()
and image() are static. I didn't notice that the code executes in a
single place -- I read it as an "ateach" instead of a nested "for". If
the code were changed to use an "ateach" (or an initializer -- see below),
my comments would still apply.
Josh, I think you meant changing
Array.makeVar[Complex](regionS, ((p):Point) => new Complex ())
to
Array.makeVar[Complex](distS, ((p):Point) => new Complex ())
rather than the Dist.makeBlock() call.
Using the initializer is a good idea, but again, if real() and image() are
not static calls, they'll be interpreted as this.real() and this.image(),
and thus result in a BadPlaceException from places other than 0. Come to
think of it, the same will happen if Mc is an instance field.
Finally, "imaginary" in complex numbers is usually abbreviated to "imag"
(without the trailing "e"). I think "image" is just a homophone.
Igor
--
Igor Peshansky (note the spelling change!)
IBM T.J. Watson Research Center
XJ: No More Pain for XML's Gain (http://www.research.ibm.com/xj/)
X10: Parallel Productivity and Performance (http://x10.sf.net/)
Josh Milthorpe <josh.miltho...@anu.edu.au>
08/28/2009 09:29 PM
To
Jim LaGrone <jlagr...@cs.uh.edu>
cc
Igor Peshansky/Watson/i...@ibmus, x10-users@lists.sourceforge.net
Subject
Re: [X10-users] Distributions and arrays
Igor Peshansky wrote:
This is slightly different code, but I realized later that this is
executed first. But this throws a similar x10.lang.BadPlaceException.
And yes, S is a member of State class. I have not yet attempted this
with C++ backend but will do so.
/* class members */
var regionS: Region{rank==2};
var distS: Dist{rank==2};
var S: Array[Complex](distS);
/* code from method */
val real = tmp_fs_real;//Rail
val image = tmp_fs_image;//Rail
regionS = [0..N-1, 0..Mc-1];
distS = Dist.makeBlock(regionS,0);
S = Array.makeVar[Complex](regionS, ((p):Point) => new Complex ());
val S = this.S;
finish{
for ( p in distS.places()) async (p){
for( (i1,j1) in regionS ){
S(i1,j1).real = real(i1*Mc+j1);
S(i1,j1).image = image(i1*Mc+j1);
}
}
}
this.S = S;
Let me guess: real() and image() are instance methods of the State
class... Again, you're attempting to invoke methods on an object
that lives in place 0 from other places, which won't work.
Basically, the rule of thumb is: you cannot access state or invoke
instance methods via a remote reference.
If real() and image() could be made static, the above code should
work.
Igor
Hi Jim,
if I correctly understand the intent, you're trying to initialize each
element of S : Array[Complex] based on some values in the real and image
rails. However I don't think that's what the above code achieves.
The outer loop iterates over all places included in distS. The inner loop
iterates over every point in the region of S. Because distS is a block
distribution over all places, this means that an activity will be started
at each place, and each activity will attempt to assign to every element
of S - including those elements that reside at other places. I suspect
that this is the cause of the BPE.
(As an aside, S is not distributed, because distS is not used in the
initializer. To make it distributed, use
distS = Dist.makeBlock(distS,0);
rather than
distS = Dist.makeBlock(regionS,0);
)
You could initialize S in a single statement using an Array initializer
function, something like:
S = Array.make[Complex](distS, (val (i,j): Point) => new Complex
(real(i*Mc+j), image(i*Mc+j)));
BTW is the use of "image" rather than "imaginary" an American vs. British
thing? I haven't heard this term used before.
Cheers,
Josh
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
X10-users mailing list
X10-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/x10-users
