Skip to content

Fix QM/MM switching funciton #446

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
lohedges merged 1 commit into from
May 29, 2026
Merged

Fix QM/MM switching funciton #446

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
25 changes: 19 additions & 6 deletions wrapper/Convert/SireOpenMM/pyqm.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -720,6 +720,7 @@ double PyQMForceImpl::computeForce(
QVector<double> charges_unscaled;
QVector<double> min_dists;
QVector<Vector> nearest_qm_vecs;
QVector<int> nearest_qm_atom_idxs;

// If we are using electrostatic embedding, the work out the MM point charges and
// build the neighbour list.
Expand Down Expand Up @@ -766,17 +767,21 @@ double PyQMForceImpl::computeForce(
// Find the minimum distance to any QM atom.
double min_dist = std::numeric_limits<double>::max();
Vector nearest_qm_vec;
int nearest_qm_atom_idx = -1;

// Loop over all of the QM atoms.
for (const auto &qm_vec : xyz_qm_vec)
for (int qm_j = 0; qm_j < xyz_qm_vec.size(); ++qm_j)
{
const auto &qm_vec = xyz_qm_vec[qm_j];

// Work out the distance between the current MM atom and QM atom.
const auto dist = space.calcDist(mm_vec, qm_vec);

if (dist < min_dist)
{
min_dist = dist;
nearest_qm_vec = qm_vec;
nearest_qm_atom_idx = qm_atoms[qm_j];
}

// The current MM atom is within the neighbour list cutoff.
Expand All @@ -799,6 +804,7 @@ double PyQMForceImpl::computeForce(
charges_unscaled.append(q);
min_dists.append(min_dist);
nearest_qm_vecs.append(nearest_qm_vec);
nearest_qm_atom_idxs.append(nearest_qm_atom_idx);

// Scale charge by switching function.
charges_mm.append(q * switching_function(min_dist));
Expand All @@ -822,14 +828,17 @@ double PyQMForceImpl::computeForce(
// Find the minimum distance to any QM atom.
double min_dist = std::numeric_limits<double>::max();
Vector nearest_qm_vec;
int nearest_qm_atom_idx = -1;

for (const auto &qm_vec : xyz_qm_vec)
for (int qm_j = 0; qm_j < xyz_qm_vec.size(); ++qm_j)
{
const auto &qm_vec = xyz_qm_vec[qm_j];
const auto dist = space.calcDist(mm_vec, qm_vec);
if (dist < min_dist)
{
min_dist = dist;
nearest_qm_vec = qm_vec;
nearest_qm_atom_idx = qm_atoms[qm_j];
}
}

Expand All @@ -843,6 +852,7 @@ double PyQMForceImpl::computeForce(
charges_unscaled.append(q);
min_dists.append(min_dist);
nearest_qm_vecs.append(nearest_qm_vec);
nearest_qm_atom_idxs.append(nearest_qm_atom_idx);

// Scale charge by switching function.
charges_mm.append(q * switching_function(min_dist));
Expand Down Expand Up @@ -1027,10 +1037,13 @@ double PyQMForceImpl::computeForce(
// multiply by 10 to convert kJ/mol/Å → kJ/mol/nm (OpenMM units).
const double correction =
-dE_dcharges_mm[i] * 10.0 * charges_unscaled[i] * dsdr;
forces[idx] += lambda * OpenMM::Vec3(
correction * r_hat[0],
correction * r_hat[1],
correction * r_hat[2]);
const OpenMM::Vec3 f_corr(correction * r_hat[0],
correction * r_hat[1],
correction * r_hat[2]);
// Apply to MM atom.
forces[idx] += lambda * f_corr;
// Apply equal and opposite force to the nearest QM atom (Newton's 3rd law).
forces[nearest_qm_atom_idxs[i]] -= lambda * f_corr;
}
}
}
Expand Down
25 changes: 19 additions & 6 deletions wrapper/Convert/SireOpenMM/torchqm.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -537,6 +537,7 @@ double TorchQMForceImpl::computeForce(
QVector<double> charges_unscaled;
QVector<double> min_dists;
QVector<Vector> nearest_qm_vecs;
QVector<int> nearest_qm_atom_idxs;

// If we are using electrostatic embedding, the work out the MM point charges and
// build the neighbour list.
Expand Down Expand Up @@ -583,17 +584,21 @@ double TorchQMForceImpl::computeForce(
// Find the minimum distance to any QM atom.
double min_dist = std::numeric_limits<double>::max();
Vector nearest_qm_vec;
int nearest_qm_atom_idx = -1;

// Loop over all of the QM atoms.
for (const auto &qm_vec : xyz_qm_vec)
for (int qm_j = 0; qm_j < xyz_qm_vec.size(); ++qm_j)
{
const auto &qm_vec = xyz_qm_vec[qm_j];

// Work out the distance between the current MM atom and QM atom.
const auto dist = space.calcDist(mm_vec, qm_vec);

if (dist < min_dist)
{
min_dist = dist;
nearest_qm_vec = qm_vec;
nearest_qm_atom_idx = qm_atoms[qm_j];
}

// The current MM atom is within the neighbour list cutoff.
Expand All @@ -618,6 +623,7 @@ double TorchQMForceImpl::computeForce(
charges_unscaled.append(q);
min_dists.append(min_dist);
nearest_qm_vecs.append(nearest_qm_vec);
nearest_qm_atom_idxs.append(nearest_qm_atom_idx);

// Scale charge by switching function.
charges_mm.append(q * switching_function(min_dist));
Expand All @@ -641,14 +647,17 @@ double TorchQMForceImpl::computeForce(
// Find the minimum distance to any QM atom.
double min_dist = std::numeric_limits<double>::max();
Vector nearest_qm_vec;
int nearest_qm_atom_idx = -1;

for (const auto &qm_vec : xyz_qm_vec)
for (int qm_j = 0; qm_j < xyz_qm_vec.size(); ++qm_j)
{
const auto &qm_vec = xyz_qm_vec[qm_j];
const auto dist = space.calcDist(mm_vec, qm_vec);
if (dist < min_dist)
{
min_dist = dist;
nearest_qm_vec = qm_vec;
nearest_qm_atom_idx = qm_atoms[qm_j];
}
}

Expand All @@ -664,6 +673,7 @@ double TorchQMForceImpl::computeForce(
charges_unscaled.append(q);
min_dists.append(min_dist);
nearest_qm_vecs.append(nearest_qm_vec);
nearest_qm_atom_idxs.append(nearest_qm_atom_idx);

// Scale charge by switching function.
charges_mm.append(q * switching_function(min_dist));
Expand Down Expand Up @@ -944,10 +954,13 @@ double TorchQMForceImpl::computeForce(
// dE_dq is in Hartree/e; dsdr is in 1/Å; multiply by HARTREE_TO_KJ_MOL*10
// to convert to kJ/mol/nm, matching the units of forces_qm/forces_mm.
const double correction = -dE_dq[i].item<float>() * HARTREE_TO_KJ_MOL * 10.0 * charges_unscaled[i] * dsdr;
forces[idx] += lambda * OpenMM::Vec3(
correction * r_hat[0],
correction * r_hat[1],
correction * r_hat[2]);
const OpenMM::Vec3 f_corr(correction * r_hat[0],
correction * r_hat[1],
correction * r_hat[2]);
// Apply to MM atom.
forces[idx] += lambda * f_corr;
// Apply equal and opposite force to the nearest QM atom (Newton's 3rd law).
forces[nearest_qm_atom_idxs[i]] -= lambda * f_corr;
}
}
}
Expand Down