Skip to content

[femutils] Remove usage of CsrFormatMatrixView:rowsNbColumn() #350

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
grospelliergilles merged 1 commit into from
Feb 17, 2026
Merged

[femutils] Remove usage of CsrFormatMatrixView:rowsNbColumn() #350

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 7 additions & 3 deletions femutils/CsrFormatMatrixView.h
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,11 +1,11 @@
// -*- tab-width: 2; indent-tabs-mode: nil; coding: utf-8-with-signature -*-
//-----------------------------------------------------------------------------
// Copyright 2000-2025 CEA (www.cea.fr) IFPEN (www.ifpenergiesnouvelles.com)
// Copyright 2000-2026 CEA (www.cea.fr) IFPEN (www.ifpenergiesnouvelles.com)
// See the top-level COPYRIGHT file for details.
// SPDX-License-Identifier: Apache-2.0
//-----------------------------------------------------------------------------
/*---------------------------------------------------------------------------*/
/* CsrFormatMatrixView.h (C) 2022-2025 */
/* CsrFormatMatrixView.h (C) 2022-2026 */
/* */
/* View of a Matrix with CSR format. */
/*---------------------------------------------------------------------------*/
Expand Down Expand Up @@ -154,7 +154,6 @@ class CsrFormatMatrixView
public:

[[nodiscard]] constexpr ARCCORE_HOST_DEVICE Span<const Int32> rows() const { return m_matrix_rows; }
[[nodiscard]] constexpr ARCCORE_HOST_DEVICE Span<const Int32> rowsNbColumn() const { return m_matrix_rows_nb_column; }
[[nodiscard]] constexpr ARCCORE_HOST_DEVICE Span<const Int32> columns() const { return m_matrix_columns; }
[[nodiscard]] constexpr ARCCORE_HOST_DEVICE Span<Real> values() const { return m_values; }

Expand All @@ -166,6 +165,11 @@ class CsrFormatMatrixView
[[nodiscard]] constexpr ARCCORE_HOST_DEVICE Int32 nbValue() const { return m_values.size(); }

[[nodiscard]] constexpr ARCCORE_HOST_DEVICE Int32 row(Int32 index) const { return m_matrix_rows[index]; }
//! Number of column for the row \a row
[[nodiscard]] constexpr ARCCORE_HOST_DEVICE Int32 nbColumnForRow(Int32 row) const
{
return m_matrix_rows_nb_column[row];
}

//! Local index of the column for the given RowColumnIndex \a rc_index
[[nodiscard]] constexpr ARCCORE_HOST_DEVICE Int32 column(CsrRowColumnIndex rc_index) const { return m_matrix_columns[rc_index]; }
Expand Down
32 changes: 19 additions & 13 deletions femutils/HypreDoFLinearSystem.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,11 +1,11 @@
// -*- tab-width: 2; indent-tabs-mode: nil; coding: utf-8-with-signature -*-
//-----------------------------------------------------------------------------
// Copyright 2000-2025 CEA (www.cea.fr) IFPEN (www.ifpenergiesnouvelles.com)
// Copyright 2000-2026 CEA (www.cea.fr) IFPEN (www.ifpenergiesnouvelles.com)
// See the top-level COPYRIGHT file for details.
// SPDX-License-Identifier: Apache-2.0
//-----------------------------------------------------------------------------
/*---------------------------------------------------------------------------*/
/* HypreDoFLinearSystem.cc (C) 2022-2025 */
/* HypreDoFLinearSystem.cc (C) 2022-2026 */
/* */
/* Linear system: Matrix A + Vector x + Vector b for Ax=b. */
/*---------------------------------------------------------------------------*/
Expand All @@ -32,7 +32,8 @@
#include <arcane/core/Timer.h>

#include <arcane/accelerator/VariableViews.h>
#include <arcane/accelerator/core/Runner.h>
#include <arcane/accelerator/NumArrayViews.h>
//#include <arcane/accelerator/core/Runner.h>
#include <arcane/accelerator/core/Memory.h>
#include <arcane/accelerator/core/DeviceMemoryInfo.h>

Expand Down Expand Up @@ -352,10 +353,10 @@ solve()
const Int32 nb_local_row = rows_index_span.size();

CSRFormatView csr_view = this->getCSRValues();
const Int32 nb_row = csr_view.nbRow();
if (do_debug_print) {
info() << "ROWS_INDEX=" << rows_index_span;
info() << "ROWS=" << csr_view.rows();
info() << "ROWS_NB_COLUMNS=" << csr_view.rowsNbColumn();
info() << "COLUMNS=" << csr_view.columns();
info() << "VALUE=" << csr_view.values();
}
Expand All @@ -374,7 +375,7 @@ solve()
}
}

int* rows_nb_column_data = const_cast<int*>(csr_view.rowsNbColumn().data());
//int* rows_nb_column_data = const_cast<int*>(csr_view.rowsNbColumn().data());

Real m1 = platform::getRealTime();
hypreCheck("IJMatrixSetObjectType",HYPRE_IJMatrixSetObjectType(ij_A, HYPRE_PARCSR));
Expand Down Expand Up @@ -414,7 +415,7 @@ solve()
if (do_debug_print) {
ENUMERATE_ (DoF, idof, dof_family->allItems()) {
DoF dof = *idof;
Int32 nb_col = csr_view.rowsNbColumn()[idof.index()];
Int32 nb_col = csr_view.nbColumnForRow(idof.index());
Int32 row_csr_index = csr_view.rows()[idof.index()];
info() << "DoF dof=" << ItemPrinter(dof) << " nb_col=" << nb_col << " row_csr_index=" << row_csr_index
<< " global_row=" << rows_index_span[idof.index()];
Expand Down Expand Up @@ -450,7 +451,6 @@ solve()
DoF dof = *idof;
if (!dof.isOwn())
continue;
Int32 nb_col = csr_view.rowsNbColumn()[idof.index()];
m_parallel_rows_index[index] = rows_index_span[idof.index()];
++index;
}
Expand All @@ -459,6 +459,7 @@ solve()
// Prefetch the memory to the Device to make sure we are using
// Device memory and not host memory when using UVM
NumArray<Int32, MDDim1> na_rows_nb_column_data(mem_ressource);
na_rows_nb_column_data.resize(nb_row);
NumArray<Int32, MDDim1> na_rows_index(mem_ressource);
NumArray<Int32, MDDim1> na_columns_index(mem_ressource);
NumArray<Real, MDDim1> na_matrix_values(mem_ressource);
Expand All @@ -474,35 +475,40 @@ solve()

if (is_use_device) {
info() << "[Hypre-Info] Prefetching memory";
q.prefetchMemory(Accelerator::MemoryPrefetchArgs(ConstMemoryView(csr_view.rowsNbColumn())).addAsync());
q.prefetchMemory(Accelerator::MemoryPrefetchArgs(ConstMemoryView(rows_index_span)).addAsync());
q.prefetchMemory(Accelerator::MemoryPrefetchArgs(ConstMemoryView(columns_index_span)).addAsync());
q.prefetchMemory(Accelerator::MemoryPrefetchArgs(ConstMemoryView(matrix_values)).addAsync());

VariableUtils::prefetchVariableAsync(rhs_variable, &q);
VariableUtils::prefetchVariableAsync(dof_variable, &q);
}
Span<const Int32> rows_nb_column_span = csr_view.rowsNbColumn();
//Span<const Int32> rows_nb_column_span = csr_view.rowsNbColumn();
if (is_use_device && is_use_device_memory) {
_doCopy(na_rows_nb_column_data, rows_nb_column_span, &q);
_doCopy(na_rows_index, rows_index_span, &q);
_doCopy(na_columns_index, columns_index_span, &q);
_doCopy(na_matrix_values, matrix_values, &q);

rows_nb_column_data = na_rows_nb_column_data.to1DSpan().data();
rows_index_data = na_rows_index.to1DSpan().data();
columns_index_data = na_columns_index.to1DSpan().data();
matrix_values_data = na_matrix_values.to1DSpan().data();

q.barrier();
}

// Fill the array containing the number of columns per row
{
auto command = makeCommand(q);
auto rows_nb_column_data = viewOut(command, na_rows_nb_column_data);
command << RUNCOMMAND_LOOP1(i, nb_row)
{
rows_nb_column_data[i] = csr_view.nbColumnForRow(i);
};
}
{
Timer::Action ta1(tstat, "HypreLinearSystemBuildMatrix");
// GPU pointers; efficient in large chunks //
HYPRE_IJMatrixSetValues(ij_A,
nb_local_row,
rows_nb_column_data,
na_rows_nb_column_data.to1DSpan().data(),
rows_index_data,
columns_index_data,
matrix_values_data);
Expand Down