Merge pull request #5906 from akva2/janitoring_const_correctness

Janitoring: const correctness
This commit is contained in:
Bård Skaflestad 2025-01-23 09:06:18 +01:00 committed by GitHub
commit 0507c87e39
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
30 changed files with 222 additions and 152 deletions

View File

@ -509,8 +509,8 @@ private:
if (!onUpperBoundary_(pos))
return false;
Scalar xInject[] = { 0.25, 0.75 };
Scalar injectLen[] = { 0.1, 0.1 };
const Scalar xInject[] = { 0.25, 0.75 };
const Scalar injectLen[] = { 0.1, 0.1 };
for (unsigned i = 0; i < sizeof(xInject) / sizeof(Scalar); ++i) {
if (xInject[i] - injectLen[i] / 2 < lambda
&& lambda < xInject[i] + injectLen[i] / 2)

View File

@ -28,12 +28,14 @@
#ifndef EWOMS_DISPERSION_MODULE_HH
#define EWOMS_DISPERSION_MODULE_HH
#include <dune/common/fvector.hh>
#include <opm/models/common/multiphasebaseproperties.hh>
#include <opm/models/discretization/common/fvbaseproperties.hh>
#include <opm/material/common/MathToolbox.hpp>
#include <opm/material/common/Valgrind.hpp>
#include <dune/common/fvector.hh>
#include <stdexcept>
namespace Opm {
@ -371,7 +373,7 @@ protected:
for (unsigned i = 0; i < phaseIdxs.size(); ++i) {
normVelocityCell_[i] = 0;
}
for (auto& velocityInfo : velocityInfos) {
for (const auto& velocityInfo : velocityInfos) {
for (unsigned i = 0; i < phaseIdxs.size(); ++i) {
if (FluidSystem::phaseIsActive(phaseIdxs[i])) {
normVelocityCell_[phaseIdxs[i]] = max( normVelocityCell_[phaseIdxs[i]],

View File

@ -136,11 +136,12 @@ BlackoilAquiferModel<TypeTag>::endTimeStep()
{
using NumAq = AquiferNumerical<TypeTag>;
for (auto& aquifer : this->aquifers) {
for (const auto& aquifer : this->aquifers) {
aquifer->endTimeStep();
NumAq* num = dynamic_cast<NumAq*>(aquifer.get());
if (num)
const NumAq* num = dynamic_cast<const NumAq*>(aquifer.get());
if (num) {
this->simulator_.vanguard().grid().comm().barrier();
}
}
}

View File

@ -81,7 +81,7 @@ void mergeParallelLogFiles(std::string_view output_dir,
enableLoggingFalloutWarning));
}
void handleExtraConvergenceOutput(SimulatorReport& report,
void handleExtraConvergenceOutput(const SimulatorReport& report,
std::string_view option,
std::string_view optionName,
std::string_view output_dir,

View File

@ -36,7 +36,7 @@ void mergeParallelLogFiles(std::string_view output_dir,
std::string_view deck_filename,
bool enableLoggingFalloutWarning);
void handleExtraConvergenceOutput(SimulatorReport& report,
void handleExtraConvergenceOutput(const SimulatorReport& report,
std::string_view option,
std::string_view optionName,
std::string_view output_dir,

View File

@ -23,6 +23,8 @@
#ifndef OPM_GENERIC_THRESHOLD_PRESSURE_IMPL_HPP
#define OPM_GENERIC_THRESHOLD_PRESSURE_IMPL_HPP
#include <opm/simulators/flow/GenericThresholdPressure.hpp>
#include <dune/grid/common/mcmgmapper.hh>
#include <dune/grid/common/rangegenerators.hh>
@ -34,8 +36,6 @@
#include <opm/input/eclipse/EclipseState/SimulationConfig/SimulationConfig.hpp>
#include <opm/input/eclipse/EclipseState/SimulationConfig/ThresholdPressure.hpp>
#include <opm/simulators/flow/GenericThresholdPressure.hpp>
#include <fmt/format.h>
#include <algorithm>
@ -73,10 +73,11 @@ thresholdPressure(int elem1Idx, int elem2Idx) const
int fault1Idx = lookUpCartesianData_(elem1Idx, cartElemFaultIdx_);
int fault2Idx = lookUpCartesianData_(elem2Idx, cartElemFaultIdx_);
if (fault1Idx != -1 && fault1Idx == fault2Idx)
if (fault1Idx != -1 && fault1Idx == fault2Idx) {
// inside a fault there's no threshold pressure, even accross EQUIL
// regions.
return 0.0;
}
if (fault1Idx != fault2Idx) {
// TODO: which value if a cell is part of multiple faults? we take
// the maximum here.
@ -90,8 +91,9 @@ thresholdPressure(int elem1Idx, int elem2Idx) const
auto equilRegion1Idx = elemEquilRegion_[elem1Idx];
auto equilRegion2Idx = elemEquilRegion_[elem2Idx];
if (equilRegion1Idx == equilRegion2Idx)
if (equilRegion1Idx == equilRegion2Idx) {
return 0.0;
}
return thpres_[equilRegion1Idx*numEquilRegions_ + equilRegion2Idx];
}
@ -103,8 +105,9 @@ finishInit()
const auto& simConfig = eclState_.getSimulationConfig();
enableThresholdPressure_ = simConfig.useThresholdPressure();
if (!enableThresholdPressure_)
if (!enableThresholdPressure_) {
return;
}
numEquilRegions_ = eclState_.getTableManager().getEqldims().getNumEquilRegions();
const decltype(numEquilRegions_) maxRegions =
@ -129,16 +132,18 @@ finishInit()
}
// internalize the data specified using the EQLNUM keyword
elemEquilRegion_ = lookUpData_.template assignFieldPropsIntOnLeaf<short unsigned int>(eclState_.fieldProps(),
"EQLNUM", true);
elemEquilRegion_ = lookUpData_.
template assignFieldPropsIntOnLeaf<short unsigned int>(eclState_.fieldProps(),
"EQLNUM", true);
/*
If this is a restart run the ThresholdPressure object will be active,
and already properly initialized with numerical values from the restart.
Done using GenericThresholdPressure::setFromRestart() in EclWriter::beginRestart().
*/
if (simConfig.getThresholdPressure().restart())
if (simConfig.getThresholdPressure().restart()) {
return;
}
// allocate the array which specifies the threshold pressures
thpres_.resize(numEquilRegions_*numEquilRegions_, 0.0);
@ -156,10 +161,12 @@ applyExplicitThresholdPressures_()
// intersection in the grid
for (const auto& elem : elements(gridView_, Dune::Partitions::interior)) {
for (const auto& intersection : intersections(gridView_, elem)) {
if (intersection.boundary())
if (intersection.boundary()) {
continue; // ignore boundary intersections for now (TODO?)
else if (!intersection.neighbor()) //processor boundary but not domain boundary
}
else if (!intersection.neighbor()) { // processor boundary but not domain boundary
continue;
}
const auto& inside = intersection.inside();
const auto& outside = intersection.outside();
@ -189,8 +196,9 @@ applyExplicitThresholdPressures_()
}
// apply threshold pressures across faults
if (thpres.ftSize() > 0)
if (thpres.ftSize() > 0) {
configureThpresft_();
}
}
template<class Grid, class GridView, class ElementMapper, class Scalar>
@ -207,14 +215,16 @@ configureThpresft_()
int numCartesianElem = eclState_.getInputGrid().getCartesianSize();
thpresftValues_.resize(numFaults, -1.0);
cartElemFaultIdx_.resize(numCartesianElem, -1);
for (std::size_t faultIdx = 0; faultIdx < faults.size(); faultIdx++) {
auto& fault = faults.getFault(faultIdx);
for (std::size_t faultIdx = 0; faultIdx < faults.size(); ++faultIdx) {
const auto& fault = faults.getFault(faultIdx);
thpresftValues_[faultIdx] = thpres.getThresholdPressureFault(faultIdx);
for (const FaultFace& face : fault)
for (const FaultFace& face : fault) {
// "face" is a misnomer because the object describes a set of cell
// indices, but we go with the conventions of the parser here...
for (std::size_t cartElemIdx : face)
for (std::size_t cartElemIdx : face) {
cartElemFaultIdx_[cartElemIdx] = faultIdx;
}
}
}
}
@ -223,8 +233,9 @@ std::vector<Scalar>
GenericThresholdPressure<Grid,GridView,ElementMapper,Scalar>::
getRestartVector() const
{
if (!enableThresholdPressure_)
if (!enableThresholdPressure_) {
return {};
}
return this->thpres_;
}
@ -243,8 +254,9 @@ void
GenericThresholdPressure<Grid,GridView,ElementMapper,Scalar>::
logPressures()
{
if (!enableThresholdPressure_)
if (!enableThresholdPressure_) {
return;
}
auto lineFormat = [this](unsigned i, unsigned j, double val)
{
@ -274,7 +286,8 @@ logPressures()
if (thpres.hasRegionBarrier(i, j)) {
if (thpres.hasThresholdPressure(i, j)) {
str += lineFormat(i, j, thpres.getThresholdPressure(j, i));
} else {
}
else {
std::size_t idx = (j - 1) * numEquilRegions_ + (i - 1);
str += lineFormat(i, j, this->thpresDefault_[idx]);
}

View File

@ -190,7 +190,7 @@ Opm::InterRegFlowMap::getInterRegFlows() const
auto maps = std::vector<data::InterRegFlowMap>{};
maps.reserve(this->regionMaps_.size());
for (auto& regionMap : this->regionMaps_) {
for (const auto& regionMap : this->regionMaps_) {
maps.push_back(regionMap.getInterRegFlows());
}
@ -203,7 +203,7 @@ Opm::InterRegFlowMap::getLocalMaxRegionID() const
auto maxLocalRegionID = std::vector<std::size_t>{};
maxLocalRegionID.reserve(this->regionMaps_.size());
for (auto& regionMap : this->regionMaps_) {
for (const auto& regionMap : this->regionMaps_) {
maxLocalRegionID.push_back(regionMap.getLocalMaxRegionID());
}

View File

@ -711,7 +711,7 @@ public:
if (!problem.model().linearizer().getFlowsInfo().empty()) {
const auto& flowsInf = problem.model().linearizer().getFlowsInfo();
auto flowsInfos = flowsInf[globalDofIdx];
for (auto& flowsInfo : flowsInfos) {
for (const auto& flowsInfo : flowsInfos) {
if (flowsInfo.faceId >= 0) {
if (!this->flows_[flowsInfo.faceId][gasCompIdx].empty()) {
this->flows_[flowsInfo.faceId][gasCompIdx][globalDofIdx]
@ -750,7 +750,7 @@ public:
if (!problem.model().linearizer().getFloresInfo().empty()) {
const auto& floresInf = problem.model().linearizer().getFloresInfo();
auto floresInfos =floresInf[globalDofIdx];
for (auto& floresInfo : floresInfos) {
for (const auto& floresInfo : floresInfos) {
if (floresInfo.faceId >= 0) {
if (!this->flores_[floresInfo.faceId][gasCompIdx].empty()) {
this->flores_[floresInfo.faceId][gasCompIdx][globalDofIdx]

View File

@ -244,7 +244,7 @@ public:
#ifdef RESERVOIR_COUPLING_ENABLED
// NOTE: The argc and argv will be used when launching a slave process
void init(SimulatorTimer &timer, int argc, char** argv)
void init(const SimulatorTimer& timer, int argc, char** argv)
{
auto slave_mode = Parameters::Get<Parameters::Slave>();
if (slave_mode) {
@ -269,7 +269,7 @@ public:
}
}
#else
void init(SimulatorTimer &timer)
void init(const SimulatorTimer& timer)
{
#endif
simulator_.setEpisodeIndex(-1);

View File

@ -616,7 +616,7 @@ protected:
if (elem.partitionType() != Dune::InteriorEntity)
{
// Dirichlet boundary conditions needed for the parallel matrix
for (auto& tr : tbatch) {
for (const auto& tr : tbatch) {
if (tr.numTracer() != 0) {
(*tr.mat)[I][I][0][0] = 1.;
(*tr.mat)[I][I][1][1] = 1.;
@ -892,7 +892,7 @@ protected:
TracerBatch(int phaseIdx = 0) : phaseIdx_(phaseIdx) {}
int numTracer() const {return idx_.size(); }
int numTracer() const { return idx_.size(); }
void addTracer(const int idx, const TV & concentration)
{

View File

@ -61,7 +61,7 @@ void sortRow(int *colIndices, int *data, int left, int right)
// LUMat->nnzValues[ik] = LUMat->nnzValues[ik] - (pivot * LUMat->nnzValues[jk]) in ilu decomposition
// a = a - (b * c)
template<class Scalar>
void blockMultSub(Scalar* a, Scalar* b, Scalar* c, unsigned int block_size)
void blockMultSub(Scalar* a, const Scalar* b, const Scalar* c, unsigned int block_size)
{
for (unsigned int row = 0; row < block_size; row++) {
for (unsigned int col = 0; col < block_size; col++) {
@ -76,7 +76,8 @@ void blockMultSub(Scalar* a, Scalar* b, Scalar* c, unsigned int block_size)
/*Perform a 3x3 matrix-matrix multiplicationj on two blocks*/
template<class Scalar>
void blockMult(Scalar* mat1, Scalar* mat2, Scalar* resMat, unsigned int block_size)
void blockMult(const Scalar* mat1, const Scalar* mat2,
Scalar* resMat, unsigned int block_size)
{
for (unsigned int row = 0; row < block_size; row++) {
for (unsigned int col = 0; col < block_size; col++) {
@ -90,8 +91,8 @@ void blockMult(Scalar* mat1, Scalar* mat2, Scalar* resMat, unsigned int block_si
}
#define INSTANTIATE_TYPE(T) \
template void blockMultSub(T*, T*, T*, unsigned int); \
template void blockMult(T*, T*, T*, unsigned int);
template void blockMultSub(T*, const T*, const T*, unsigned int); \
template void blockMult(const T*, const T*, T*, unsigned int);
INSTANTIATE_TYPE(double)

View File

@ -111,7 +111,8 @@ void sortRow(int* colIndices, int* data, int left, int right);
/// \param[in] c input block
/// \param[in] block_size size of block
template<class Scalar>
void blockMultSub(Scalar* a, Scalar* b, Scalar* c, unsigned int block_size);
void blockMultSub(Scalar* a, const Scalar* b,
const Scalar* c, unsigned int block_size);
/// Perform a matrix-matrix multiplication on two blocks
/// resMat = mat1 * mat2

View File

@ -23,28 +23,28 @@
#include <opm/simulators/linalg/gpubridge/Reorder.hpp>
#include <vector>
#include <cassert>
#include <vector>
namespace Opm
{
namespace Accelerator
{
namespace Opm::Accelerator {
/* Check is operations on a node in the matrix can be started
* A node can only be started if all nodes that it depends on during sequential execution have already completed.*/
* A node can only be started if all nodes that it depends on during sequential
* execution have already completed.*/
bool canBeStarted(const int rowIndex, const int *rowPointers, const int *colIndices, const std::vector<bool>& doneRows) {
bool canBeStarted(const int rowIndex,
const int* rowPointers,
const int* colIndices,
const std::vector<bool>& doneRows)
{
bool canStart = !doneRows[rowIndex];
int i, thisDependency;
if (canStart) {
for (i = rowPointers[rowIndex]; i < rowPointers[rowIndex + 1]; i++) {
thisDependency = colIndices[i];
for (int i = rowPointers[rowIndex]; i < rowPointers[rowIndex + 1]; ++i) {
int thisDependency = colIndices[i];
// Only dependencies on rows that should execute before the current one are relevant
if (thisDependency >= rowIndex)
if (thisDependency >= rowIndex) {
break;
}
// Check if dependency has been resolved
if (!doneRows[thisDependency]) {
return false;
@ -55,14 +55,23 @@ bool canBeStarted(const int rowIndex, const int *rowPointers, const int *colIndi
}
/*
* The level scheduling of a non-symmetric, blocked matrix requires access to a CSC encoding and a CSR encoding of the sparsity pattern of the input matrix.
* The level scheduling of a non-symmetric, blocked matrix requires access to a CSC
* encoding and a CSR encoding of the sparsity pattern of the input matrix.
* This function is based on a standard level scheduling algorithm, like the one described in:
* "Iterative methods for Sparse Linear Systems" by Yousef Saad in section 11.6.3
*/
void findLevelScheduling(int *CSRColIndices, int *CSRRowPointers, int *CSCRowIndices, int *CSCColPointers, int Nb, int *numColors, int *toOrder, int* fromOrder, std::vector<int>& rowsPerColor) {
int activeRowIndex = 0, colorEnd, nextActiveRowIndex = 0;
int thisRow;
void findLevelScheduling(const int* CSRColIndices,
const int* CSRRowPointers,
const int* CSCRowIndices,
const int* CSCColPointers,
int Nb,
int* numColors,
int* toOrder,
int* fromOrder,
std::vector<int>& rowsPerColor)
{
int activeRowIndex = 0, nextActiveRowIndex = 0;
std::vector<bool> doneRows(Nb, false);
std::vector <int> rowsToStart;
@ -70,23 +79,27 @@ void findLevelScheduling(int *CSRColIndices, int *CSRRowPointers, int *CSCRowInd
assert(rowsPerColor.empty());
// find starting rows: rows that are independent from all rows that come before them.
for (thisRow = 0; thisRow < Nb; thisRow++) {
int thisRow;
for (thisRow = 0; thisRow < Nb; ++thisRow) {
if (canBeStarted(thisRow, CSCColPointers, CSCRowIndices, doneRows)) {
fromOrder[nextActiveRowIndex] = thisRow;
toOrder[thisRow] = nextActiveRowIndex;
nextActiveRowIndex++;
++nextActiveRowIndex;
}
}
// 'do' compute on all active rows
for (colorEnd = 0; colorEnd < nextActiveRowIndex; colorEnd++) {
int colorEnd;
for (colorEnd = 0; colorEnd < nextActiveRowIndex; ++colorEnd) {
doneRows[fromOrder[colorEnd]] = true;
}
rowsPerColor.emplace_back(nextActiveRowIndex - activeRowIndex);
while (colorEnd < Nb) {
// Go over all rows active from the last color, and check which of their neighbours can be activated this color
for (; activeRowIndex < colorEnd; activeRowIndex++) {
// Go over all rows active from the last color, and check which of
// their neighbours can be activated this color
for (; activeRowIndex < colorEnd; ++activeRowIndex) {
thisRow = fromOrder[activeRowIndex];
for (int i = CSCColPointers[thisRow]; i < CSCColPointers[thisRow + 1]; i++) {
@ -104,7 +117,7 @@ void findLevelScheduling(int *CSRColIndices, int *CSRRowPointers, int *CSCRowInd
doneRows[thisRow] = true;
fromOrder[nextActiveRowIndex] = thisRow;
toOrder[thisRow] = nextActiveRowIndex;
nextActiveRowIndex++;
++nextActiveRowIndex;
}
}
rowsToStart.clear();
@ -115,10 +128,13 @@ void findLevelScheduling(int *CSRColIndices, int *CSRRowPointers, int *CSCRowInd
*numColors = rowsPerColor.size();
}
// based on the scipy package from python, scipy/sparse/sparsetools/csr.h on github
void csrPatternToCsc(int *CSRColIndices, int *CSRRowPointers, int *CSCRowIndices, int *CSCColPointers, int Nb) {
void csrPatternToCsc(const int* CSRColIndices,
const int* CSRRowPointers,
int* CSCRowIndices,
int* CSCColPointers,
int Nb)
{
int nnz = CSRRowPointers[Nb];
// compute number of nnzs per column
@ -141,7 +157,7 @@ void csrPatternToCsc(int *CSRColIndices, int *CSRRowPointers, int *CSCRowIndices
int col = CSRColIndices[j];
int dest = CSCColPointers[col];
CSCRowIndices[dest] = row;
CSCColPointers[col]++;
++CSCColPointers[col];
}
}
@ -152,6 +168,4 @@ void csrPatternToCsc(int *CSRColIndices, int *CSRRowPointers, int *CSCRowIndices
}
}
} // namespace Accelerator
} // namespace Opm
} // namespace Opm::Accelerator

View File

@ -22,10 +22,7 @@
#include <vector>
namespace Opm
{
namespace Accelerator
{
namespace Opm::Accelerator {
/// Determine whether all rows that a certain row depends on are done already
/// \param[in] rowIndex index of the row that needs to be checked for
@ -33,7 +30,10 @@ namespace Accelerator
/// \param[in] colIndices column indices of the matrix that the row is in
/// \param[in] doneRows array that for each row lists whether it is done or not
/// \return true iff all dependencies are done and if the result itself was not done yet
bool canBeStarted(const int rowIndex, const int *rowPointers, const int *colIndices, const std::vector<bool>& doneRows);
bool canBeStarted(const int rowIndex,
const int* rowPointers,
const int* colIndices,
const std::vector<bool>& doneRows);
/// Find a level scheduling reordering for an input matrix
/// The toOrder and fromOrder arrays must be allocated already
@ -46,7 +46,15 @@ bool canBeStarted(const int rowIndex, const int *rowPointers, const int *colIn
/// \param[out] toOrder the reorder pattern that was found, which lists for each index in the original order, to which index in the new order it should be moved
/// \param[out] fromOrder the reorder pattern that was found, which lists for each index in the new order, from which index in the original order it was moved
/// \param[out] rowsPerColor for each color, an array of all rowIndices in that color, this function uses emplace_back() to fill
void findLevelScheduling(int *CSRColIndices, int *CSRRowPointers, int *CSCRowIndices, int *CSCColPointers, int Nb, int *numColors, int *toOrder, int* fromOrder, std::vector<int>& rowsPerColor);
void findLevelScheduling(const int* CSRColIndices,
const int* CSRRowPointers,
const int* CSCRowIndices,
const int* CSCColPointers,
int Nb,
int* numColors,
int* toOrder,
int* fromOrder,
std::vector<int>& rowsPerColor);
/// Convert a sparsity pattern stored in the CSR format to the CSC format
/// CSCRowIndices and CSCColPointers arrays must be allocated already
@ -56,9 +64,12 @@ void findLevelScheduling(int *CSRColIndices, int *CSRRowPointers, int *CSCRowInd
/// \param[inout] CSCRowIndices row indices of the result CSC representation of the pattern
/// \param[inout] CSCColPointers column pointers of the result CSC representation of the pattern
/// \param[in] Nb number of blockrows in the matrix
void csrPatternToCsc(int *CSRColIndices, int *CSRRowPointers, int *CSCRowIndices, int *CSCColPointers, int Nb);
void csrPatternToCsc(const int* CSRColIndices,
const int* CSRRowPointers,
int* CSCRowIndices,
int* CSCColPointers,
int Nb);
} // namespace Accelerator
} // namespace Opm
} // namespace Opm::Accelerator
#endif
#endif

View File

@ -163,7 +163,7 @@ void amgclSolverBackend<Scalar,block_size>::initialize(int Nb_, int nnzbs)
template<class Scalar, unsigned int block_size>
void amgclSolverBackend<Scalar,block_size>::
convert_sparsity_pattern(int* rows, int* cols)
convert_sparsity_pattern(const int* rows, const int* cols)
{
Timer t;
const unsigned int bs = block_size;
@ -193,7 +193,7 @@ convert_sparsity_pattern(int* rows, int* cols)
template<class Scalar, unsigned int block_size>
void amgclSolverBackend<Scalar,block_size>::
convert_data(Scalar* vals, int* rows)
convert_data(const Scalar* vals, const int* rows)
{
Timer t;
const unsigned int bs = block_size;

View File

@ -105,12 +105,12 @@ private:
/// Convert the BCSR sparsity pattern to a CSR one
/// \param[in] rows array of rowPointers, contains N/dim+1 values
/// \param[in] cols array of columnIndices, contains nnz values
void convert_sparsity_pattern(int *rows, int *cols);
void convert_sparsity_pattern(const int *rows, const int *cols);
/// Convert the BCSR nonzero data to a CSR format
/// \param[in] vals array of nonzeroes, each block is stored row-wise and contiguous, contains nnz values
/// \param[in] rows array of rowPointers, contains N/dim+1 values
void convert_data(Scalar* vals, int* rows);
void convert_data(const Scalar* vals, const int* rows);
/// Solve linear system
/// \param[in] b pointer to b vector

View File

@ -194,12 +194,13 @@ create_preconditioner(BlockedMatrix<Scalar>* mat, BlockedMatrix<Scalar>* jacMat)
{
const unsigned int bs = block_size;
auto *matToDecompose = jacMat ? jacMat : mat;
const auto* matToDecompose = jacMat ? jacMat : mat;
bool use_multithreading = true;
#if HAVE_OPENMP
if (omp_get_max_threads() == 1)
if (omp_get_max_threads() == 1) {
use_multithreading = false;
}
#endif
if (jacMat && use_multithreading) {

View File

@ -18,14 +18,19 @@
*/
#ifndef OPM_GPUISTL_GPUOWNEROVERLAPCOPY_HPP
#define OPM_GPUISTL_GPUOWNEROVERLAPCOPY_HPP
#include <dune/istl/owneroverlapcopy.hh>
#include <opm/simulators/linalg/gpuistl/GpuVector.hpp>
#include <mpi.h>
#include <memory>
#include <mutex>
#include <opm/simulators/linalg/gpuistl/GpuVector.hpp>
#include <vector>
namespace Opm::gpuistl
{
namespace Opm::gpuistl {
/**
* @brief GPUSender is a wrapper class for classes which will implement copOwnerToAll
* This is implemented with the intention of creating communicators with generic GPUSender
@ -119,10 +124,11 @@ public:
explicit GPUObliviousMPISender(const OwnerOverlapCopyCommunicationType& cpuOwnerOverlapCopy)
: GPUSender<field_type, OwnerOverlapCopyCommunicationType>(cpuOwnerOverlapCopy)
{
}
{
}
void copyOwnerToAll(const X& source, X& dest) const override {
void copyOwnerToAll(const X& source, X& dest) const override
{
// TODO: [perf] Can we reduce copying from the GPU here?
// TODO: [perf] Maybe create a global buffer instead?
auto sourceAsDuneVector = source.template asDuneBlockVector<block_size>();
@ -179,7 +185,6 @@ public:
void copyOwnerToAll(const X& source, X& dest) const override
{
OPM_ERROR_IF(&source != &dest, "The provided GpuVectors' address did not match"); // In this context, source == dest!!!
std::call_once(this->m_initializedIndices, [&]() { initIndexSet(); });
@ -198,9 +203,9 @@ public:
{
size_t i = 0;
for(const_iterator info = m_messageInformation.begin(); info != end; ++info, ++i) {
for (const_iterator info = m_messageInformation.begin(); info != end; ++info, ++i) {
processMap[i]=info->first;
if(info->second.second.m_size) {
if (info->second.second.m_size) {
MPI_Irecv(m_GPURecvBuf->data()+info->second.second.m_start,
detail::to_int(info->second.second.m_size),
MPI_BYTE,
@ -209,16 +214,17 @@ public:
this->m_cpuOwnerOverlapCopy.communicator(),
&recvRequests[i]);
numberOfRealRecvRequests += 1;
} else {
recvRequests[i]=MPI_REQUEST_NULL;
}
else {
recvRequests[i] = MPI_REQUEST_NULL;
}
}
}
{
size_t i = 0;
for(const_iterator info = m_messageInformation.begin(); info != end; ++info, ++i) {
if(info->second.first.m_size) {
for (const_iterator info = m_messageInformation.begin(); info != end; ++info, ++i) {
if (info->second.first.m_size) {
MPI_Issend(m_GPUSendBuf->data()+info->second.first.m_start,
detail::to_int(info->second.first.m_size),
MPI_BYTE,
@ -227,24 +233,28 @@ public:
this->m_cpuOwnerOverlapCopy.communicator(),
&sendRequests[i]);
} else {
sendRequests[i]=MPI_REQUEST_NULL;
sendRequests[i] = MPI_REQUEST_NULL;
}
}
}
int finished = MPI_UNDEFINED;
MPI_Status status;
for(size_t i = 0; i < numberOfRealRecvRequests; i++) {
for (size_t i = 0; i < numberOfRealRecvRequests; i++) {
status.MPI_ERROR=MPI_SUCCESS;
MPI_Waitany(m_messageInformation.size(), recvRequests.data(), &finished, &status);
if(status.MPI_ERROR!=MPI_SUCCESS) {
OPM_THROW(std::runtime_error, fmt::format("MPI_Error occurred while rank {} received a message from rank {}", rank, processMap[finished]));
if (status.MPI_ERROR!=MPI_SUCCESS) {
OPM_THROW(std::runtime_error,
fmt::format("MPI_Error occurred while rank {} received a message from rank {}",
rank, processMap[finished]));
}
}
MPI_Status recvStatus;
for(size_t i = 0; i < m_messageInformation.size(); i++) {
if(MPI_SUCCESS!=MPI_Wait(&sendRequests[i], &recvStatus)) {
OPM_THROW(std::runtime_error, fmt::format("MPI_Error occurred while rank {} sent a message from rank {}", rank, processMap[finished]));
for (size_t i = 0; i < m_messageInformation.size(); i++) {
if (MPI_SUCCESS != MPI_Wait(&sendRequests[i], &recvStatus)) {
OPM_THROW(std::runtime_error,
fmt::format("MPI_Error occurred while rank {} sent a message from rank {}",
rank, processMap[finished]));
}
}
// ...End of MPI stuff
@ -275,24 +285,25 @@ private:
void buildCommPairIdxs() const
{
auto &ri = this->m_cpuOwnerOverlapCopy.remoteIndices();
const auto& ri = this->m_cpuOwnerOverlapCopy.remoteIndices();
std::vector<int> commpairIndicesCopyOnCPU;
std::vector<int> commpairIndicesOwnerCPU;
for(auto process : ri) {
for (const auto& process : ri) {
m_im[process.first] = std::pair(std::vector<int>(), std::vector<int>());
for(int send = 0; send < 2; ++send) {
for (int send = 0; send < 2; ++send) {
auto remoteEnd = send ? process.second.first->end()
: process.second.second->end();
auto remote = send ? process.second.first->begin()
: process.second.second->begin();
while(remote != remoteEnd) {
while (remote != remoteEnd) {
if (send ? (remote->localIndexPair().local().attribute() == 1)
: (remote->attribute() == 1)) {
if (send) {
m_im[process.first].first.push_back(remote->localIndexPair().local().local());
} else {
}
else {
m_im[process.first].second.push_back(remote->localIndexPair().local().local());
}
}
@ -317,13 +328,13 @@ private:
recvBufIdx * block_size,
noRecv * block_size * sizeof(field_type)))));
for(int x = 0; x < noSend; x++) {
for(int bs = 0; bs < block_size; bs++) {
for (int x = 0; x < noSend; x++) {
for (int bs = 0; bs < block_size; bs++) {
commpairIndicesOwnerCPU.push_back(it->second.first[x] * block_size + bs);
}
}
for(int x = 0; x < noRecv; x++) {
for(int bs = 0; bs < block_size; bs++) {
for (int x = 0; x < noRecv; x++) {
for (int bs = 0; bs < block_size; bs++) {
commpairIndicesCopyOnCPU.push_back(it->second.second[x] * block_size + bs);
}
}
@ -385,9 +396,12 @@ class GpuOwnerOverlapCopy
public:
using X = GpuVector<field_type>;
explicit GpuOwnerOverlapCopy(std::shared_ptr<GPUSender<field_type, OwnerOverlapCopyCommunicationType>> sender) : m_sender(sender){}
explicit GpuOwnerOverlapCopy(std::shared_ptr<GPUSender<field_type, OwnerOverlapCopyCommunicationType>> sender)
: m_sender(sender)
{}
void copyOwnerToAll(const X& source, X& dest) const {
void copyOwnerToAll(const X& source, X& dest) const
{
m_sender->copyOwnerToAll(source, dest);
}
@ -409,5 +423,7 @@ public:
private:
std::shared_ptr<GPUSender<field_type, OwnerOverlapCopyCommunicationType>> m_sender;
};
} // namespace Opm::gpuistl
#endif

View File

@ -6,17 +6,22 @@
// NOTE: This file is a modified version of dune/istl/paamg/twolevelmethod.hh from
// dune-istl release 2.6.0. Modifications have been kept as minimal as possible.
#include <tuple>
#include<dune/istl/operators.hh>
#include <dune/istl/operators.hh>
//#include "amg.hh"
//#include"galerkin.hh"
#include<dune/istl/paamg/amg.hh>
#include<dune/istl/paamg/galerkin.hh>
#include<dune/istl/solver.hh>
#include <dune/istl/paamg/amg.hh>
#include <dune/istl/paamg/galerkin.hh>
#include <dune/istl/solver.hh>
#include<dune/common/unused.hh>
#include<dune/common/version.hh>
#include <dune/common/unused.hh>
#include <dune/common/version.hh>
#include <algorithm>
#include <cstddef>
#include <iostream>
#include <memory>
#include <tuple>
#include <vector>
/**
* @addtogroup ISTL_PAAMG
@ -189,13 +194,8 @@ public:
ParallelInformation pinfo;
std::size_t aggregates = coarsen(renumberer, pinfo, pg, vm,*aggregatesMap_, noAggregates, true);
std::vector<bool>& visited=excluded;
typedef std::vector<bool>::iterator Iterator;
for(Iterator iter= visited.begin(), end=visited.end();
iter != end; ++iter)
*iter=false;
std::fill(excluded.begin(), excluded.end(), false);
matrix_.reset(productBuilder.build(mg, vm,
SequentialInformation(),
*aggregatesMap_,

View File

@ -542,7 +542,7 @@ extern "C" {
if (cells.empty()) { return; }
const auto first = parts[cells.front()];
for (auto& cell : cells) {
for (const auto& cell : cells) {
parts[cell] = first;
}
}

View File

@ -259,7 +259,7 @@ template<class Scalar> class WellContributions;
// add source from wells to the reservoir matrix
void addReservoirSourceTerms(GlobalEqVector& residual,
std::vector<typename SparseMatrixAdapter::MatrixBlock*>& diagMatAddress) const;
const std::vector<typename SparseMatrixAdapter::MatrixBlock*>& diagMatAddress) const;
// called at the beginning of a report step
void beginReportStep(const int time_step);

View File

@ -1906,7 +1906,7 @@ getMaxWellConnections() const
const auto possibleFutureConnectionSetIt = possibleFutureConnections.find(well.name());
if (possibleFutureConnectionSetIt != possibleFutureConnections.end()) {
for (auto& global_index : possibleFutureConnectionSetIt->second) {
for (const auto& global_index : possibleFutureConnectionSetIt->second) {
int compressed_idx = compressedIndexForInterior(global_index);
if (compressed_idx >= 0) { // Ignore connections in inactive/remote cells.
compressed_well_perforations.push_back(compressed_idx);

View File

@ -433,7 +433,7 @@ namespace Opm {
// have proper multi-phase rates proportional to rates at bhp zero.
// This is done only for producers, as injectors will only have a single
// nonzero phase anyway.
for (auto& well : well_container_) {
for (const auto& well : well_container_) {
const bool zero_target = well->stoppedOrZeroRateTarget(simulator_, this->wellState(), local_deferredLogger);
if (well->isProducer() && !zero_target) {
well->updateWellStateRates(simulator_, this->wellState(), local_deferredLogger);
@ -441,7 +441,7 @@ namespace Opm {
}
}
for (auto& well : well_container_) {
for (const auto& well : well_container_) {
if (well->isVFPActive(local_deferredLogger)){
well->setPrevSurfaceRates(this->wellState(), this->prevWellState());
}
@ -478,7 +478,7 @@ namespace Opm {
auto exc_type = ExceptionType::NONE;
// update gpmaint targets
if (this->schedule_[reportStepIdx].has_gpmaint()) {
for (auto& calculator : regionalAveragePressureCalculator_) {
for (const auto& calculator : regionalAveragePressureCalculator_) {
calculator.second->template defineState<ElementContext>(simulator_);
}
const double dt = simulator_.timeStepSize();
@ -1641,7 +1641,7 @@ namespace Opm {
template <typename TypeTag>
void BlackoilWellModel<TypeTag>::
addReservoirSourceTerms(GlobalEqVector& residual,
std::vector<typename SparseMatrixAdapter::MatrixBlock*>& diagMatAddress) const
const std::vector<typename SparseMatrixAdapter::MatrixBlock*>& diagMatAddress) const
{
// NB this loop may write multiple times to the same element
// if a cell is perforated by more than one well, so it should
@ -2011,7 +2011,10 @@ namespace Opm {
const int iterationIdx,
DeferredLogger& deferred_logger)
{
this->updateAndCommunicateGroupData(reportStepIdx, iterationIdx, param_.nupcol_group_rate_tolerance_, deferred_logger);
this->updateAndCommunicateGroupData(reportStepIdx,
iterationIdx,
param_.nupcol_group_rate_tolerance_,
deferred_logger);
// updateWellStateWithTarget might throw for multisegment wells hence we
// have a parallel try catch here to thrown on all processes.
@ -2019,15 +2022,20 @@ namespace Opm {
// if a well or group change control it affects all wells that are under the same group
for (const auto& well : well_container_) {
// We only want to update wells under group-control here
auto& ws = this->wellState().well(well->indexOfWell());
if (ws.production_cmode == Well::ProducerCMode::GRUP || ws.injection_cmode == Well::InjectorCMode::GRUP) {
const auto& ws = this->wellState().well(well->indexOfWell());
if (ws.production_cmode == Well::ProducerCMode::GRUP ||
ws.injection_cmode == Well::InjectorCMode::GRUP)
{
well->updateWellStateWithTarget(simulator_, this->groupState(),
this->wellState(), deferred_logger);
}
}
OPM_END_PARALLEL_TRY_CATCH("BlackoilWellModel::updateAndCommunicate failed: ",
simulator_.gridView().comm())
this->updateAndCommunicateGroupData(reportStepIdx, iterationIdx, param_.nupcol_group_rate_tolerance_, deferred_logger);
this->updateAndCommunicateGroupData(reportStepIdx,
iterationIdx,
param_.nupcol_group_rate_tolerance_,
deferred_logger);
}
template<typename TypeTag>

View File

@ -1590,7 +1590,7 @@ template<class Scalar>
void GasLiftSingleWellGeneric<Scalar>::
updateWellStateAlqFixedValue_(const GasLiftWell& well)
{
auto& max_alq_optional = well.max_rate();
const auto& max_alq_optional = well.max_rate();
if (max_alq_optional) {
// According to WLIFTOPT, item 3:
// If item 2 is NO, then item 3 is regarded as the fixed

View File

@ -215,7 +215,7 @@ void
GasLiftSingleWell<TypeTag>::
setAlqMaxRate_(const GasLiftWell& well)
{
auto& max_alq_optional = well.max_rate();
const auto& max_alq_optional = well.max_rate();
if (max_alq_optional) {
// NOTE: To prevent extrapolation of the VFP tables, any value
// entered here must not exceed the largest ALQ value in the well's VFP table.

View File

@ -474,7 +474,7 @@ recalculateGradientAndUpdateData_(GradPairItr& grad_itr,
// only applies to wells in the well_state_map (i.e. wells on this rank)
// the grads and other grads are synchronized later
if(this->stage1_wells_.count(name) > 0) {
GasLiftSingleWell &gs_well = *(this->stage1_wells_.at(name).get());
const GasLiftSingleWell& gs_well = *(this->stage1_wells_.at(name).get());
{
auto grad = calcIncOrDecGrad_(name, gs_well, gr_name_dont_limit, increase);
if (grad) {

View File

@ -121,7 +121,7 @@ public:
// Create a function that calls some function
// for all the individual data items to simplify
// the further code.
auto iterateContainer = [](auto& container, auto& func) {
auto iterateContainer = [](auto& container, const auto& func) {
for (auto& x : container) {
func(x.second);
}

View File

@ -81,7 +81,7 @@ public:
// Now add the cells of the possible future connections
const auto possibleFutureConnectionSetIt = possibleFutureConnections.find(well.name());
if (possibleFutureConnectionSetIt != possibleFutureConnections.end()) {
for (auto& global_index : possibleFutureConnectionSetIt->second) {
for (const auto& global_index : possibleFutureConnectionSetIt->second) {
int compressed_idx = model_.compressedIndexForInterior(global_index);
if (compressed_idx >= 0) { // Ignore connections in inactive/remote cells.
wellCells.push_back(compressed_idx);
@ -126,7 +126,7 @@ public:
}
}
void postSolveDomain(GlobalEqVector& deltaX, const Domain& domain)
void postSolveDomain(const GlobalEqVector& deltaX, const Domain& domain)
{
model_.recoverWellSolutionAndUpdateWellStateDomain(deltaX, domain);
}

View File

@ -476,7 +476,7 @@ namespace Opm
deferred_logger.info(msg);
// also reopen completions
for (auto& completion : this->well_ecl_.getCompletions()) {
for (const auto& completion : this->well_ecl_.getCompletions()) {
if (!welltest_state_temp.completion_is_closed(this->name(), completion.first))
well_test_state.open_completion(this->name(), completion.first);
}

View File

@ -116,7 +116,8 @@ createBridge(const boost::property_tree::ptree& prm, std::unique_ptr<Opm::GpuBri
template <int bz>
Dune::BlockVector<Dune::FieldVector<double, bz>>
testOpenclSolver(Opm::GpuBridge<Matrix<bz>, Vector<bz>, bz>& bridge, Matrix<bz>& matrix, Vector<bz>& rhs)
testOpenclSolver(Opm::GpuBridge<Matrix<bz>, Vector<bz>, bz>& bridge,
const Matrix<bz>& matrix, Vector<bz>& rhs)
{
Dune::InverseOperatorResult result;
Vector<bz> x(rhs.size());
@ -131,7 +132,8 @@ testOpenclSolver(Opm::GpuBridge<Matrix<bz>, Vector<bz>, bz>& bridge, Matrix<bz>&
template <int bz>
Dune::BlockVector<Dune::FieldVector<double, bz>>
testOpenclSolverJacobi(Opm::GpuBridge<Matrix<bz>, Vector<bz>, bz>& bridge, Matrix<bz>& matrix, Vector<bz>& rhs)
testOpenclSolverJacobi(Opm::GpuBridge<Matrix<bz>, Vector<bz>, bz>& bridge,
const Matrix<bz>& matrix, Vector<bz>& rhs)
{
Dune::InverseOperatorResult result;
Vector<bz> x(rhs.size());