A code issue about force plugin about openmm

[peastman]

I don't see the force being calculated twice. Each time through the loop it prints two lines (one with force_old and one with force_new). And NUM_PARTICLECORR is clearly 2, since there are lines with index=0 and index=1. So there should be four lines printed for each step, which is what it shows. Am I missing something?

[yongyouhe]

I don't see the force being calculated twice. Each time through the loop it prints two lines (one with force_old and one with force_new). And NUM_PARTICLECORR is clearly 2, since there are lines with index=0 and index=1. So there should be four lines printed for each step, which is what it shows. Am I missing something?

Maybe I didn't express the question clearly.
There are two identical step 1110 being output. Is it normal to output twice? Then why are force_old (0.0 and -13.897074 for index=0) and force_new (0.021701 and -13.875373 for index=0) different each time they are output?
The twice outputs imply that the execute() function of this forcekernel was executed twice. Where should this function be called separately? I can only think of one place: each step LanveginIntegrator is executed. context->calcForcesAndEnergy(true, false, getIntegrationForceGroups());

[peastman]

Can you show your code that executes it? If the force is being evaluated twice, there's generally a reason for it.

[yongyouhe]

Can you show your code that executes it? If the force is being evaluated twice, there's generally a reason for it.

Sorry to reply so late. I had found the problem -- I made a simple mistake when writing the simulation run script. For testing purposes, I ran the simulation for one step and then output the energy using getState() each time. So, it is the function getState() that leads to the second output of Step 1110.

Thanks for your reply!
Best regards.