I am calculating lipid acyl chain order using SCD tool of MEMBPlugin from 20,000 frame trajectory, total time of trajectory is 100 ns.
My concern/question is that it is very slow to calculate the SCD order from the tool as it took two days to calculate order parameter for 2000 frames, so for me it will take 20 days to calculate 20,000 frames. I have even stripped water from my trajectory to make the calculations fast but it didnt affect anything.
Kindly let me know is there any way I can calculate SCD order parameter fast than this, or am I making any mistake to use the tool.
Hoping to get a positive response
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The SCD is a computational expensive calculation. We know that Tcl scripting maybe is not fast enough and that our implementation maybe is not fully optimized. For the next release, we will optimize the performance of this tool.
At the moment, I would recommend you striding your trajectory to a number of frames that gives you a reasonable standard error/ deviation and a good representation of your system (maybe a timestep of 1.0-0.1 ns).
Another way to speed up computation could be writing on your disk the trajectory with only the atoms and frames needed for computation (e.g. all your lipids). Then, you can perform the computations from the new trajectory file (this would optimize disk access).
Writing this new trajectory already wrapped and using a blank wrap command (by deleting the default command in text box and leaving it empty) during calculation could be also a good idea.
Finally, splitting your trajectory in several files, and running VMD parallel instances at the same time for different parts of the trajectory could be also a solution. For this, you need to have available several CPUs. Once you obtained the output files, you can process them with a statistics/ math external program.
I hope this helps.
Last edit: Ismael Rodriguez Espigares 2015-07-08
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thank you very much for your response, I will try what you have proposed and get back to you as soon as possible. I have so many other questions also related to MEMBPLUGIN tool. and will also ask them.. .
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear all
I am calculating lipid acyl chain order using SCD tool of MEMBPlugin from 20,000 frame trajectory, total time of trajectory is 100 ns.
My concern/question is that it is very slow to calculate the SCD order from the tool as it took two days to calculate order parameter for 2000 frames, so for me it will take 20 days to calculate 20,000 frames. I have even stripped water from my trajectory to make the calculations fast but it didnt affect anything.
Kindly let me know is there any way I can calculate SCD order parameter fast than this, or am I making any mistake to use the tool.
Hoping to get a positive response
The SCD is a computational expensive calculation. We know that Tcl scripting maybe is not fast enough and that our implementation maybe is not fully optimized. For the next release, we will optimize the performance of this tool.
At the moment, I would recommend you striding your trajectory to a number of frames that gives you a reasonable standard error/ deviation and a good representation of your system (maybe a timestep of 1.0-0.1 ns).
Another way to speed up computation could be writing on your disk the trajectory with only the atoms and frames needed for computation (e.g. all your lipids). Then, you can perform the computations from the new trajectory file (this would optimize disk access).
Writing this new trajectory already wrapped and using a blank wrap command (by deleting the default command in text box and leaving it empty) during calculation could be also a good idea.
Finally, splitting your trajectory in several files, and running VMD parallel instances at the same time for different parts of the trajectory could be also a solution. For this, you need to have available several CPUs. Once you obtained the output files, you can process them with a statistics/ math external program.
I hope this helps.
Last edit: Ismael Rodriguez Espigares 2015-07-08
Thank you very much for your response, I will try what you have proposed and get back to you as soon as possible. I have so many other questions also related to MEMBPLUGIN tool. and will also ask them.. .