Skip to content
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

OMP parallelize sim_trans_impl #417

Open
wants to merge 4 commits into
base: main
Choose a base branch
from
Open

OMP parallelize sim_trans_impl #417

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
0907786
OMP parallelize sim_trans_impl
brianz98 Sep 24, 2024
ad58c1b
Manual reduction for gcc error (clang worked fine)
brianz98 Sep 25, 2024
67f9a88
Merge branch 'main' into simtrans
brianz98 Nov 15, 2024
3187da1
Merge branch 'main' into simtrans
brianz98 Nov 19, 2024
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
1 change: 0 additions & 1 deletion forte/sparse_ci/sparse_operator.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,7 +151,6 @@ SparseOperator commutator(const SparseOperator& lhs, const SparseOperator& rhs)
return C;
}


void SparseOperatorList::add_term_from_str(std::string str, sparse_scalar_t coefficient,
bool allow_reordering) {
auto [sqop, phase] = make_sq_operator_string(str, allow_reordering);
Expand Down
42 changes: 31 additions & 11 deletions forte/sparse_ci/sparse_operator_sim_trans.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,8 +25,15 @@
*
* @END LICENSE
*/

#include <numeric>
#include "sparse_operator.h"
#ifdef _OPENMP
#include <omp.h>
#else
#define omp_get_max_threads() 1
#define omp_get_thread_num() 0
#define omp_get_num_threads() 1
#endif

namespace forte {

Expand Down Expand Up @@ -219,12 +226,23 @@ void sim_trans_impl(SparseOperator& O, const SQOperatorString& T_op,
auto Td_n = Td_op.number_component();
auto Td_nn = Td_op.non_number_component();

SparseOperator T;
int nthreads = omp_get_max_threads();

std::vector<SparseOperator> T(nthreads);
sparse_scalar_t c1, c2;
Determinant nn_Op_cre;
Determinant nn_Op_ann;

for (const auto& [O_op, O_c] : O.elements()) {
SparseOperatorList O_list;
for (const auto& [sqop, c] : O.elements()) {
O_list.add(sqop, c);
}

#pragma omp parallel private(c1, c2) num_threads(nthreads)
{
int ithread = omp_get_thread_num();
#pragma omp for schedule(dynamic, 10)
for (const auto& [O_op, O_c] : O_list.elements()) {
const auto O_T_commutator_type = commutator_type(O_op, T_op);
const auto O_Td_commutator_type = commutator_type(O_op, Td_op);
// if both commutators are zero, then we can skip this term
Expand Down Expand Up @@ -289,18 +307,18 @@ void sim_trans_impl(SparseOperator& O, const SQOperatorString& T_op,
for (const auto& [cOT_op, cOT_c] : cOT) {
// + c1 [O, T]
if (do_c1) {
T.add(cOT_op, cOT_c * c1);
T.at(ithread).add(cOT_op, cOT_c * c1);
}
if (do_c2) {
auto ccOTT = commutator_fast(cOT_op, T_op);
for (const auto& [ccOTT_op, ccOTT_c] : ccOTT) {
// + c2 [[O, T], T]
T.add(ccOTT_op, ccOTT_c * cOT_c * c2);
T.at(ithread).add(ccOTT_op, ccOTT_c * cOT_c * c2);
}
auto ccOTTd = commutator_fast(cOT_op, Td_op);
for (const auto& [ccOTTd_op, ccOTTd_c] : ccOTTd) {
// sigma * c2 [[O, T], T^dagger]
T.add(ccOTTd_op, sigma * ccOTTd_c * cOT_c * c2);
T.at(ithread).add(ccOTTd_op, sigma * ccOTTd_c * cOT_c * c2);
}
}
}
Expand All @@ -311,27 +329,29 @@ void sim_trans_impl(SparseOperator& O, const SQOperatorString& T_op,
for (const auto& [cOTd_op, cOTd_c] : cOTd) {
// sigma * c1 [O, T^dagger]
if (do_c1) {
T.add(cOTd_op, sigma * cOTd_c * c1);
T.at(ithread).add(cOTd_op, sigma * cOTd_c * c1);
}
if (do_c2) {
auto ccOTdT = commutator_fast(cOTd_op, T_op);
for (const auto& [ccOTdT_op, ccOTdT_c] : ccOTdT) {
// sigma * c2 [[O, T^dagger], T]
T.add(ccOTdT_op, sigma * ccOTdT_c * cOTd_c * c2);
T.at(ithread).add(ccOTdT_op, sigma * ccOTdT_c * cOTd_c * c2);
}
auto ccOTdTd = commutator_fast(cOTd_op, Td_op);
for (const auto& [ccOTdTd_op, ccOTdTd_c] : ccOTdTd) {
// +c2 [[O, T^dagger], T^dagger]
T.add(ccOTdTd_op, ccOTdTd_c * cOTd_c * c2);
T.at(ithread).add(ccOTdTd_op, ccOTdTd_c * cOTd_c * c2);
}
}
}
}
}
}
SparseOperator T_sum = std::accumulate(T.begin(), T.end(), SparseOperator());
if (add) {
O += T;
O += T_sum;
} else {
O = T;
O = T_sum;
}
}

Expand Down
Loading