Russel Winder wrote: > "Pi by Quadrature" again (sorry), but this time trying to use > parallelism via partitioning the problem using an parallel array. As a > first step I thought of using the Array[T] class -- or should that be > Array<T> the API page uses both notations, cf. > http://dist.codehaus.org/x10/xdoc/2.0.3/ > > I had thought that something along the lines of: > > val computeSlice = ( p : Point ) => { > val id : int = p(0) ; > val start : long = 1 + id * sliceSize ; > val end : long = ( id + 1 ) * sliceSize ; > var sum : double = 0.0 ; > for ( var i : long = start ; i < end ; ++i ) { > val x : double = ( i - 0.5 ) * delta ; > sum += 1.0 / ( 1.0 + x * x ) ; > } > sum > } ; > val sums = ( new Array[Double] ( numberOfTasks) ).lift > ( computeSlice ) ; > > > <<<< I am not convinced this operation labelled "lift" is actually a > "lift" operation, it is actually "map" isn't it? >>>> > Yes, this method should be called map.
------------------------------------------------------------------------------ _______________________________________________ X10-users mailing list X10-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/x10-users
