Re: Non bonding energy calculation through NAMD

From: Pratik Narain Srivastava (pratik.narain_at_gmail.com)
Date: Fri Apr 12 2019 - 14:03:43 CDT

@Peter thank you. Thats a nice trick.

On Sat, 13 Apr, 2019, 12:31 AM Peter Freddolino, <petefred_at_umich.edu> wrote:

> As an alternative, you can start namdenergy with the -debug flag set
> (either via the command line or GUI interface), but then kill the namd job
> that it spawns, and then namd yourself instead. Behind the scenes it just
> writes some temporary files to your working directory that can be run using
> whatever combination of namd commands you wish.
> Best,
> Peter
>
> On Fri, Apr 12, 2019 at 2:51 PM Gumbart, JC <gumbart_at_physics.gatech.edu>
> wrote:
>
>> You can modify the source code of the NAMD energy plugin. Find the line
>> where it actually runs NAMD and change it to correspond to what you want
>> (+p<number of CPUs>).
>>
>> Best,
>> JC
>>
>> On Apr 12, 2019, at 9:58 AM, Brian Radak <brian.radak_at_gmail.com> wrote:
>>
>> I haven't worked extensively with the NAMDEnergy plugin -- I would
>> consult the VMD plugin documentation or ask that question on the vmd-l
>> listserv.
>>
>> On Fri, Apr 12, 2019 at 9:24 AM varun dewaker <
>> varun_dewaker1986_at_yahoo.com> wrote:
>>
>>> Dear Sir
>>> 1. Using NAMDEnergy plugin I am calculating Nonbond energy between
>>> selection-1 (protein) and selection-2 (ligand) from MD trajectories.
>>> 2. Under default setup NAMDEnergy plugin is using single core for this
>>> computation.
>>> 3. Following is My question:
>>> How to activate multicore processing/processors to calculate Nonbond
>>> energy between selection-1 (protein) and selection-2 (ligand) from MD
>>> trajectories in NAMDEnergy plugin?
>>>
>>> Thanks in advance.
>>>
>>> *Regards,*
>>> *Varun Kumar Dewaker*
>>> *Research Fellow*
>>> QSAR & Molecular Modelling Lab
>>> MPC Division (CSN-202)
>>> CSIR- Central Drug Research Institute
>>> Lucknow - 226031 (U.P), India
>>> *Mob:+918860194990*
>>>
>>>
>>> On Friday, 12 April, 2019, 6:21:30 pm IST, Brian Radak <
>>> brian.radak_at_gmail.com> wrote:
>>>
>>>
>>> There are numerous tutorials on this as well as discussion in the user
>>> guide -- I would recommend that you read these. The command line arguments
>>> for multiple CPUs depends on the particular binary you are using and in
>>> what context (multiple nodes, etc.), however the multicore version is the
>>> simplest and most common:
>>>
>>> $ namd2 +p<number of CPUs> <config file>
>>>
>>> On Fri, Apr 12, 2019 at 3:29 AM varun dewaker <
>>> varun_dewaker1986_at_yahoo.com> wrote:
>>>
>>> Dear Sir, CPU based calculation work fine, but it uses only single
>>> core. so it takes long time in interaction calculation . Kindly help me in
>>> sorting, how to instructs the NAMD energy program, so it uses all the cpu
>>> cores available.
>>>
>>> *Regards,*
>>> *Varun Kumar Dewaker*
>>> *Research Fellow*
>>> QSAR & Molecular Modelling Lab
>>> MPC Division (CSN-202)
>>> CSIR- Central Drug Research Institute
>>> Lucknow - 226031 (U.P), India
>>> *Mob:+918860194990*
>>>
>>>
>>> On Thursday, 11 April, 2019, 7:28:44 pm IST, Brian Radak <
>>> brian.radak_at_gmail.com> wrote:
>>>
>>>
>>> That message means just what it says - the pairInteraction keyword is
>>> not Cuda compatible. Try a CPU implementation.
>>>
>>> On Thu, Apr 11, 2019, 2:50 AM varun dewaker <
>>> varun_dewaker1986_at_remove_yahoo.com> wrote:
>>>
>>> Dear Sir
>>>
>>> I run the NAMD energy with NAMD multicore CUDA application. It says
>>> pair interaction calculation is not supported in CUDA version.
>>> What to do next?
>>>
>>> *Regards,*
>>> *Varun Kumar Dewaker*
>>> *Research Fellow*
>>> QSAR & Molecular Modelling Lab
>>> MPC Division (CSN-202)
>>> CSIR- Central Drug Research Institute
>>> Lucknow - 226031 (U.P), India
>>> *Mob:+918860194990*
>>>
>>>
>>> On Tuesday, 2 April, 2019, 7:34:23 pm IST, Giacomo Fiorin <
>>> giacomo.fiorin_at_gmail.com> wrote:
>>>
>>>
>>> Hi Varun, this calculation can be parallelized on CPU cores and/or GPU
>>> cores, as far as the NAMD executable you use supports it (i.e. you will
>>> probably want to use multicore-CUDA).
>>>
>>> To load a trajectory into NAMD directly without using VMD, take a look
>>> at coorfile:
>>>
>>> http://www.ks.uiuc.edu/Research/namd/2.13/ug/node9.html#SECTION00052200000000000000
>>>
>>> On Tue, Apr 2, 2019 at 5:02 AM varun dewaker <
>>> varun_dewaker1986_at_remove_yahoo.com> wrote:
>>>
>>> Dear NAMD users, my question is regarding the non-binding energy
>>> calculation. I want to calculate the non-binding energy from 26
>>> GB trajectory file generated by 100ns MD simulation.
>>>
>>> My first question is to confirm that this energy calculation runs only
>>> on a single processor or it is possible to run it on multicore or on GPU
>>> enabled system (since it takes a lot of time in the calculation).
>>>
>>> Another question is that it is possible to calculate non-binding energy
>>> without uploading the trajectories in the VMD.
>>> kindly suggest the syntax for this (without uploading the trajectories
>>> in VMD) since I am unable to upload all trajectories due to the limitation
>>> of RAM.
>>>
>>>
>>>
>>> *Regards,*
>>> *Varun Kumar Dewaker*
>>> *Research Fellow*
>>> QSAR & Molecular Modelling Lab
>>> MPC Division (CSN-202)
>>> CSIR- Central Drug Research Institute
>>> Lucknow - 226031 (U.P), India
>>> *Mob:+918860194990*
>>>
>>>
>>>
>>> --
>>> Giacomo Fiorin
>>> Associate Professor of Research, Temple University, Philadelphia, PA
>>> Contractor, National Institutes of Health, Bethesda, MD
>>> http://goo.gl/Q3TBQU
>>> https://github.com/giacomofiorin
>>>
>>>
>>

This archive was generated by hypermail 2.1.6 : Tue Sep 17 2019 - 23:20:45 CDT