Open euhruska opened 6 years ago
to reduce time necessary for TICA: 1) try calculating covariances on each node in parallel and the on one node calculate the TICA https://github.com/markovmodel/PyEMMA/issues/1351
2) alternative: only calculate the feature-trajectories on each node from all-atom trajectories and on one node load these feature-trajectories to calculate TICA
feature parallelization implemented
to reduce time necessary for TICA: 1) try calculating covariances on each node in parallel and the on one node calculate the TICA https://github.com/markovmodel/PyEMMA/issues/1351
2) alternative: only calculate the feature-trajectories on each node from all-atom trajectories and on one node load these feature-trajectories to calculate TICA