vtk-m/vtkm/worklet/contourtree/MergeTree.h
Kenneth Moreland bddad9b386 Remove TryExecute from filters
Now that the dispatcher does its own TryExecute, filters do not need to
do that. This change requires all worklets called by filters to be able
to execute without knowing the device a priori.
2018-10-16 15:59:53 -06:00

339 lines
14 KiB
C++

//============================================================================
// Copyright (c) Kitware, Inc.
// All rights reserved.
// See LICENSE.txt for details.
// This software is distributed WITHOUT ANY WARRANTY; without even
// the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
// PURPOSE. See the above copyright notice for more information.
//
// Copyright 2014 National Technology & Engineering Solutions of Sandia, LLC (NTESS).
// Copyright 2014 UT-Battelle, LLC.
// Copyright 2014 Los Alamos National Security.
//
// Under the terms of Contract DE-NA0003525 with NTESS,
// the U.S. Government retains certain rights in this software.
//
// Under the terms of Contract DE-AC52-06NA25396 with Los Alamos National
// Laboratory (LANL), the U.S. Government retains certain rights in
// this software.
//============================================================================
// Copyright (c) 2016, Los Alamos National Security, LLC
// All rights reserved.
//
// Copyright 2016. Los Alamos National Security, LLC.
// This software was produced under U.S. Government contract DE-AC52-06NA25396
// for Los Alamos National Laboratory (LANL), which is operated by
// Los Alamos National Security, LLC for the U.S. Department of Energy.
// The U.S. Government has rights to use, reproduce, and distribute this
// software. NEITHER THE GOVERNMENT NOR LOS ALAMOS NATIONAL SECURITY, LLC
// MAKES ANY WARRANTY, EXPRESS OR IMPLIED, OR ASSUMES ANY LIABILITY FOR THE
// USE OF THIS SOFTWARE. If software is modified to produce derivative works,
// such modified software should be clearly marked, so as not to confuse it
// with the version available from LANL.
//
// Additionally, redistribution and use in source and binary forms, with or
// without modification, are permitted provided that the following conditions
// are met:
//
// 1. Redistributions of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
// 2. Redistributions in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
// 3. Neither the name of Los Alamos National Security, LLC, Los Alamos
// National Laboratory, LANL, the U.S. Government, nor the names of its
// contributors may be used to endorse or promote products derived from
// this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY LOS ALAMOS NATIONAL SECURITY, LLC AND
// CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING,
// BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
// FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL LOS ALAMOS
// NATIONAL SECURITY, LLC OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
// INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
// BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
// USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
// THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
//============================================================================
// This code is based on the algorithm presented in the paper:
// “Parallel Peak Pruning for Scalable SMP Contour Tree Computation.”
// Hamish Carr, Gunther Weber, Christopher Sewell, and James Ahrens.
// Proceedings of the IEEE Symposium on Large Data Analysis and Visualization
// (LDAV), October 2016, Baltimore, Maryland.
//=======================================================================================
//
// COMMENTS:
//
// If we have computed the merge max & merge saddles correctly, we have substantially
// computed the merge tree already. However, it is not in the same format as we have
// previously represented it - in particular, we have yet to define all the merge arcs
// and the superarcs we have collected are not the same as before - i.e. they are already
// partially collapsed, but not according to the same rule as branch decomposition
// This unit is therefore to get the same result out as before so we can set up an
// automated crosscheck on the computation
//
// Compared to earlier versions, we have made a significant change - the merge tree
// is only computed on critical points, not on the full array. We therefore have a
// final step: to extend it to the full array. To do this, we will keep the initial
// mergeArcs array which records a maximum for each vertex, as we need the information
//
// Each maximum is now labelled with the saddle it is mapped to, or to the global min
// We therefore transfer this information back to the mergeArcs array, so that maxima
// (including saddles) are marked with the (lower) vertex that is the low end of their
// arc
// BIG CHANGE: we can actually reuse the mergeArcs array for the final merge arc, for the
// chain maximum for each (regular) point, and for the merge saddle for maxima. This is
// slightly tricky and has some extra memory traffic, but it avoids duplicating arrays
// unnecessarily
//
// Initially, mergeArcs will be set to an outbound neighbour (or self for extrema), as the
// chainMaximum array used to be.
//
// After chains are built, then it will hold *AN* accessible extremum for each vertex.
//
// During the main processing, when an extremum is assigned a saddle, it will be stored
// here. Regular points will still store pointers to an extremum.
//
// After this is done, if the mergeArc points lower/higher, it is pointing to a saddle.
// Otherwise it is pointing to an extremum.
//
// And after the final pass, it will always point to the next along superarcs.
//
//=======================================================================================
#ifndef vtkm_worklet_contourtree_mergetree_h
#define vtkm_worklet_contourtree_mergetree_h
#include <vtkm/cont/Algorithm.h>
#include <vtkm/cont/ArrayCopy.h>
#include <vtkm/cont/ArrayHandle.h>
#include <vtkm/cont/ArrayHandleConstant.h>
#include <vtkm/cont/DataSet.h>
#include <vtkm/worklet/contourtree/ChainDoubler.h>
#include <vtkm/worklet/contourtree/JoinArcConnector.h>
#include <vtkm/worklet/contourtree/JoinSuperArcFinder.h>
#include <vtkm/worklet/contourtree/PrintVectors.h>
#include <vtkm/worklet/contourtree/VertexMergeComparator.h>
//#define DEBUG_PRINT 1
//#define DEBUG_FUNCTION_ENTRY 1
//#define DEBUG_TIMING 1
namespace vtkm
{
namespace worklet
{
namespace contourtree
{
template <typename T, typename StorageType>
class MergeTree
{
public:
// original data array
const vtkm::cont::ArrayHandle<T, StorageType>& values;
// size of mesh
vtkm::Id nRows, nCols, nSlices, nVertices, nLogSteps;
// whether it is join or split tree
bool isJoinTree;
// vector of arcs representing the merge tree
vtkm::cont::ArrayHandle<vtkm::Id> mergeArcs;
// vector storing an extremum for each vertex
vtkm::cont::ArrayHandle<vtkm::Id> extrema;
// vector storing a saddle for each vertex
vtkm::cont::ArrayHandle<vtkm::Id> saddles;
// merge tree constructor
MergeTree(const vtkm::cont::ArrayHandle<T, StorageType>& Values,
vtkm::Id NRows,
vtkm::Id NCols,
vtkm::Id NSlices,
bool IsJoinTree);
// routine that does pointer-doubling in the mergeArc array
void BuildRegularChains();
// routine that computes the augmented merge tree superarcs from the merge graph
void ComputeAugmentedSuperarcs();
// routine that computes the augmented merge arcs from the superarcs
// this is separate from the previous routine because it also gets called separately
// once saddle & extrema are set for a given set of vertices, the merge arcs can be
// computed for any subset of those vertices that contains all of the critical points
void ComputeAugmentedArcs(vtkm::cont::ArrayHandle<vtkm::Id>& vertices);
// debug routine
void DebugPrint(const char* message);
};
// creates merge tree
template <typename T, typename StorageType>
MergeTree<T, StorageType>::MergeTree(const vtkm::cont::ArrayHandle<T, StorageType>& Values,
vtkm::Id NRows,
vtkm::Id NCols,
vtkm::Id NSlices,
bool IsJoinTree)
: values(Values)
, nRows(NRows)
, nCols(NCols)
, nSlices(NSlices)
, isJoinTree(IsJoinTree)
{
nVertices = nRows * nCols * nSlices;
nLogSteps = 1;
for (vtkm::Id shifter = nVertices; shifter != 0; shifter >>= 1)
nLogSteps++;
vtkm::cont::ArrayHandleConstant<vtkm::Id> nullArray(0, nVertices);
mergeArcs.Allocate(nVertices);
extrema.Allocate(nVertices);
saddles.Allocate(nVertices);
vtkm::cont::ArrayCopy(nullArray, mergeArcs);
vtkm::cont::ArrayCopy(nullArray, extrema);
vtkm::cont::ArrayCopy(nullArray, saddles);
}
// routine that does pointer-doubling in the saddles array
template <typename T, typename StorageType>
void MergeTree<T, StorageType>::BuildRegularChains()
{
#ifdef DEBUG_FUNCTION_ENTRY
std::cout << std::endl;
std::cout << "====================" << std::endl;
std::cout << "Build Regular Chains" << std::endl;
std::cout << "====================" << std::endl;
std::cout << std::endl;
#endif
// 2. Create a temporary array so that we can alternate writing between them
vtkm::cont::ArrayHandle<vtkm::Id> temporaryArcs;
temporaryArcs.Allocate(nVertices);
vtkm::cont::ArrayHandleIndex vertexIndexArray(nVertices);
ChainDoubler chainDoubler;
vtkm::worklet::DispatcherMapField<ChainDoubler> chainDoublerDispatcher(chainDoubler);
// 3. Apply pointer-doubling to build chains to maxima, rocking between two arrays
for (vtkm::Id logStep = 0; logStep < nLogSteps; logStep++)
{
chainDoublerDispatcher.Invoke(vertexIndexArray, // input
extrema); // i/o whole array
}
} // BuildRegularChains()
// routine that computes the augmented merge tree from the merge graph
template <typename T, typename StorageType>
void MergeTree<T, StorageType>::ComputeAugmentedSuperarcs()
{
#ifdef DEBUG_FUNCTION_ENTRY
std::cout << std::endl;
std::cout << "=================================" << std::endl;
std::cout << "Compute Augmented Merge Superarcs" << std::endl;
std::cout << "=================================" << std::endl;
std::cout << std::endl;
#endif
// our first step is to assign every vertex to a pseudo-extremum based on how the
// vertex ascends to a extremum, and the sequence of pruning for the extremum
// to do this, we iterate as many times as pruning occurred
// we run a little loop for each element until it finds its join superarc
// expressed as a functor.
vtkm::Id nExtrema = extrema.GetNumberOfValues();
JoinSuperArcFinder<T> joinSuperArcFinder(isJoinTree);
vtkm::worklet::DispatcherMapField<JoinSuperArcFinder<T>> joinSuperArcFinderDispatcher(
joinSuperArcFinder);
vtkm::cont::ArrayHandleIndex vertexIndexArray(nExtrema);
joinSuperArcFinderDispatcher.Invoke(vertexIndexArray, // input
values, // input (whole array)
saddles, // i/o (whole array)
extrema); // i/o (whole array)
// at the end of this, all vertices should have a pseudo-extremum in the extrema array
// and a pseudo-saddle in the saddles array
#ifdef DEBUG_PRINT
DebugPrint("Merge Superarcs Set");
#endif
} // ComputeAugmentedSuperarcs()
// routine that computes the augmented merge arcs from the superarcs
// this is separate from the previous routine because it also gets called separately
// once saddle & extrema are set for a given set of vertices, the merge arcs can be
// computed for any subset of those vertices that contains all of the critical points
template <typename T, typename StorageType>
void MergeTree<T, StorageType>::ComputeAugmentedArcs(vtkm::cont::ArrayHandle<vtkm::Id>& vertices)
{
#ifdef DEBUG_FUNCTION_ENTRY
std::cout << std::endl;
std::cout << "============================" << std::endl;
std::cout << "Compute Augmented Merge Arcs" << std::endl;
std::cout << "============================" << std::endl;
std::cout << std::endl;
#endif
// create a vector of indices for sorting
vtkm::Id nCriticalVerts = vertices.GetNumberOfValues();
vtkm::cont::ArrayHandle<vtkm::Id> vertexSorter;
vtkm::cont::ArrayCopy(vertices, vertexSorter);
// We sort by pseudo-maximum to establish the extents
vtkm::cont::Algorithm::Sort(vertexSorter,
VertexMergeComparator<T, StorageType>(values, extrema, isJoinTree));
#ifdef DEBUG_PRINT
DebugPrint("Sorting Complete");
#endif
vtkm::cont::ArrayHandleConstant<vtkm::Id> noVertArray(NO_VERTEX_ASSIGNED, nVertices);
vtkm::cont::ArrayCopy(noVertArray, mergeArcs);
vtkm::cont::ArrayHandleIndex critVertexIndexArray(nCriticalVerts);
JoinArcConnector joinArcConnector;
vtkm::worklet::DispatcherMapField<JoinArcConnector> joinArcConnectorDispatcher(joinArcConnector);
joinArcConnectorDispatcher.Invoke(critVertexIndexArray, // input
vertexSorter, // input (whole array)
extrema, // input (whole array)
saddles, // input (whole array)
mergeArcs); // output (whole array)
#ifdef DEBUG_PRINT
DebugPrint("Augmented Arcs Set");
#endif
} // ComputeAugmentedArcs()
// debug routine
template <typename T, typename StorageType>
void MergeTree<T, StorageType>::DebugPrint(const char* message)
{
std::cout << "---------------------------" << std::endl;
std::cout << std::string(message) << std::endl;
std::cout << "---------------------------" << std::endl;
std::cout << std::endl;
printLabelledBlock("Values", values, nRows * nSlices, nCols);
std::cout << std::endl;
printLabelledBlock("MergeArcs", mergeArcs, nRows, nCols);
std::cout << std::endl;
printLabelledBlock("Extrema", extrema, nRows, nCols);
std::cout << std::endl;
printLabelledBlock("Saddles", saddles, nRows, nCols);
std::cout << std::endl;
} // DebugPrint()
}
}
}
#endif