-
Notifications
You must be signed in to change notification settings - Fork 17
Open
Description
Hi,
Maybe it is a stupid question but I was reading on Velocity Verlet and how to implement it. I was wondering if the v_step_full
at line 38 of mdgrad/torchmd/solvers.py should be:
v_step_full = y[0] + v_step_half + 0.5 * a_dt * dt
Here is the link to the file + line:
Line 38 in cea2332
v_step_full = v_step_half + 0.5 * a_dt * dt |
If I'm not wrong, the algorithm should be something like this:
v[i] = v[i-1] + 0.5 * dt * a
x[i] = x[i-1] + dt * v[i]
a = f(x[i])
v[i] = v[i] + 0.5 * dt * a
best,
Lucien
Metadata
Metadata
Assignees
Labels
No labels