Kokkos Node API and Local Linear Algebra Kernels Version of the Day
Kokkos_TBBNode.hpp
00001 //@HEADER
00002 // ************************************************************************
00003 // 
00004 //          Kokkos: Node API and Parallel Node Kernels
00005 //              Copyright (2008) Sandia Corporation
00006 // 
00007 // Under the terms of Contract DE-AC04-94AL85000 with Sandia Corporation,
00008 // the U.S. Government retains certain rights in this software.
00009 // 
00010 // Redistribution and use in source and binary forms, with or without
00011 // modification, are permitted provided that the following conditions are
00012 // met:
00013 //
00014 // 1. Redistributions of source code must retain the above copyright
00015 // notice, this list of conditions and the following disclaimer.
00016 //
00017 // 2. Redistributions in binary form must reproduce the above copyright
00018 // notice, this list of conditions and the following disclaimer in the
00019 // documentation and/or other materials provided with the distribution.
00020 //
00021 // 3. Neither the name of the Corporation nor the names of the
00022 // contributors may be used to endorse or promote products derived from
00023 // this software without specific prior written permission.
00024 //
00025 // THIS SOFTWARE IS PROVIDED BY SANDIA CORPORATION "AS IS" AND ANY
00026 // EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
00027 // IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
00028 // PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL SANDIA CORPORATION OR THE
00029 // CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
00030 // EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
00031 // PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
00032 // PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
00033 // LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
00034 // NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
00035 // SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
00036 //
00037 // Questions? Contact Michael A. Heroux (maherou@sandia.gov) 
00038 // 
00039 // ************************************************************************
00040 //@HEADER
00041 
00042 #ifndef KOKKOS_TBBNODE_HPP_
00043 #define KOKKOS_TBBNODE_HPP_
00044 
00045 #include "Kokkos_StandardNodeMemoryModel.hpp"
00046 #include "Kokkos_NodeHelpers.hpp"
00047 
00048 #include <tbb/blocked_range.h>
00049 #include <tbb/parallel_for.h>
00050 #include <tbb/parallel_reduce.h>
00051 #include <tbb/task_scheduler_init.h>
00052 
00053 namespace Teuchos {
00054   // forward declarations
00055   class ParameterList;
00056 }
00057 
00058 #include <stdlib.h>
00059 
00060 namespace Kokkos {
00061 
00062   template <class WDPin>
00063   struct BlockedRangeWDP {
00064     mutable WDPin *wd;
00065     inline BlockedRangeWDP(WDPin &in) : wd(&in) {}
00066     inline void operator()(tbb::blocked_range<int> &rng) const { 
00067       for (int i=rng.begin(); i != rng.end(); ++i) wd->execute(i);
00068     }
00069   };
00070   
00071   template <class WDPin>
00072   struct BlockedRangeWDPReducer {
00073     WDPin &wd;
00074     typename WDPin::ReductionType result;
00075     BlockedRangeWDPReducer(WDPin &in) : wd(in), result(WDPin::identity()) {}
00076     BlockedRangeWDPReducer(BlockedRangeWDPReducer &in, tbb::split) : wd(in.wd) {result = wd.identity();}
00077     void operator()(tbb::blocked_range<int> &rng)
00078     { 
00079       typename WDPin::ReductionType tmpi;
00080       int end = rng.end();
00081       for (int i=rng.begin(); i != end; ++i) {
00082         tmpi = wd.generate(i);
00083         result = wd.reduce( result, tmpi );
00084       }
00085     }
00086     inline void join( const BlockedRangeWDPReducer<WDPin> &other ) {
00087       result = wd.reduce( result, other.result );
00088     }
00089   };
00090   
00094   class TBBNode : public StandardNodeMemoryModel {
00095     public:
00096   
00102       TBBNode(Teuchos::ParameterList &pl);
00103   
00106       ~TBBNode();
00107 
00112       void init(int numThreads);
00113 
00115       template <class WDP>
00116       static void parallel_for(int begin, int end, WDP wd) {
00117         BlockedRangeWDP<WDP> tbb_wd(wd);
00118         tbb::parallel_for(tbb::blocked_range<int>(begin,end), tbb_wd, tbb::auto_partitioner()); 
00119       }
00120 
00122       template <class WDP>
00123       static typename WDP::ReductionType
00124       parallel_reduce(int begin, int end, WDP wd) {
00125         BlockedRangeWDPReducer<WDP> tbb_wd(wd);
00126         tbb::parallel_reduce(tbb::blocked_range<int>(begin,end), tbb_wd, tbb::auto_partitioner());
00127         return tbb_wd.result;
00128       }
00129 
00131       inline void sync() const {};
00132   
00133     private:
00134       bool alreadyInit_;
00135       tbb::task_scheduler_init tsi_;
00136   
00137   };
00138   
00139   template <> class ArrayOfViewsHelper<TBBNode> : public ArrayOfViewsHelperTrivialImpl<TBBNode> {};
00140 
00141 } // end of Kokkos namespace
00142 
00143 #endif
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends