Re: [gmx-users] Speeding up simulation

2010-08-18 Thread Gaurav Goel 
Large cut-offs can bring artificial artifacts.

e.g. *Chemical Physics
Letters*
 Volume 406, Issues
1-3,
23 April 2005, Pages 49-53

-Gaurav

On Wed, Aug 18, 2010 at 12:04 PM, sapna sarupria
wrote:

> That might actually be the best approach speed wise. But it depends on what
> level of accuracy you wish for your calculations. What is your system like?
> May be you can have a large cut off but you will have to see what the trade
> off between speed and accuracy is.
>
>
> On Wed, Aug 18, 2010 at 10:26 AM, ms  wrote:
>
>> On 17/08/10 22:36, Justin A. Lemkul wrote:
>>
>>  There is no way to limit PME to certain ranges. It is a method for
>>> solving infinite sums. Usually PME is a poor choice for in vacuo
>>> simulations and the like, since there's nothing to be done for most of
>>> the simulation box.
>>>
>>
>> This interests me -I've used simple cut-off for in vacuo tests, but I
>> still hadn't made my mind on what electrostatic model I should truly use. Do
>> you have any suggestion?
>>
>> thanks,
>> M.
>>
>>
>>  You may want to see how others (in the literature)
>>> deal with droplet-type simulations.
>>>
>>> -Justin
>>>
>>>
 Thanks for your help.

 Sapna

 On Tue, Aug 17, 2010 at 5:29 PM, Justin A. Lemkul >>> > wrote:



 sapna sarupria wrote:

 Hi all,

 I am running (NVT) simulations of a drop of water (~5 nm in
 radius) in a big box (~25 nm) in a box and find that the
 simulations are rather slow. I am getting about 0.8 ns per day
 when a simulation of bulk system of equivalent number of waters
 will be much faster. The number of waters is ~12000. I was
 wondering if anyone can suggest methods with which I can speed
 up the simulations. I am using domain decomposition and
 optimize_fft is set to yes. PME is used for the electrostatics.


 I would suspect that the lag comes from a lot of unused PME
 calculations. During your run, PME grid points will be assigned to
 vacuum space, for which nothing needs to be done. You can check
 imbalances and performance loss in the log file.

 -Justin


 Thank you

 Sincerely
 Sapna


 -- 

 Justin A. Lemkul
 Ph.D. Candidate
 ICTAS Doctoral Scholar
 MILES-IGERT Trainee
 Department of Biochemistry
 Virginia Tech
 Blacksburg, VA
 jalemkul[at]vt.edu  | (540) 231-9080
 http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

 
 -- gmx-users mailing list gmx-users@gromacs.org
 
 http://lists.gromacs.org/mailman/listinfo/gmx-users
 Please search the archive at http://www.gromacs.org/search before
 posting!
 Please don't post (un)subscribe requests to the list. Use the www
 interface or send it to gmx-users-requ...@gromacs.org
 .
 Can't post? Read http://www.gromacs.org/mailing_lists/users.php




 --
 Sapna Sarupria
 Post-doctoral Researcher
 Princeton University
 New Jersey 08540
 U.S.A.

 Life isn't about finding yourself. Life is about creating yourself. --
 George Bernard Shaw.
 Dare to Dream


>>>
>> --
>> gmx-users mailing listgmx-users@gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at http://www.gromacs.org/search before
>> posting!
>> Please don't post (un)subscribe requests to the list. Use the www
>> interface or send it to gmx-users-requ...@gromacs.org.
>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>
>
>
>
> --
> Sapna Sarupria
> Post-doctoral Researcher
> Princeton University
> New Jersey 08540
> U.S.A.
>
> Life isn't about finding yourself. Life is about creating yourself. --
> George Bernard Shaw.
> Dare to Dream
>
> --
> gmx-users mailing listgmx-users@gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-requ...@gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>
-- 
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.grom

Re: [gmx-users] Speeding up simulation

2010-08-18 Thread sapna sarupria 
That might actually be the best approach speed wise. But it depends on what
level of accuracy you wish for your calculations. What is your system like?
May be you can have a large cut off but you will have to see what the trade
off between speed and accuracy is.

On Wed, Aug 18, 2010 at 10:26 AM, ms  wrote:

> On 17/08/10 22:36, Justin A. Lemkul wrote:
>
>  There is no way to limit PME to certain ranges. It is a method for
>> solving infinite sums. Usually PME is a poor choice for in vacuo
>> simulations and the like, since there's nothing to be done for most of
>> the simulation box.
>>
>
> This interests me -I've used simple cut-off for in vacuo tests, but I still
> hadn't made my mind on what electrostatic model I should truly use. Do you
> have any suggestion?
>
> thanks,
> M.
>
>
>  You may want to see how others (in the literature)
>> deal with droplet-type simulations.
>>
>> -Justin
>>
>>
>>> Thanks for your help.
>>>
>>> Sapna
>>>
>>> On Tue, Aug 17, 2010 at 5:29 PM, Justin A. Lemkul >> > wrote:
>>>
>>>
>>>
>>> sapna sarupria wrote:
>>>
>>> Hi all,
>>>
>>> I am running (NVT) simulations of a drop of water (~5 nm in
>>> radius) in a big box (~25 nm) in a box and find that the
>>> simulations are rather slow. I am getting about 0.8 ns per day
>>> when a simulation of bulk system of equivalent number of waters
>>> will be much faster. The number of waters is ~12000. I was
>>> wondering if anyone can suggest methods with which I can speed
>>> up the simulations. I am using domain decomposition and
>>> optimize_fft is set to yes. PME is used for the electrostatics.
>>>
>>>
>>> I would suspect that the lag comes from a lot of unused PME
>>> calculations. During your run, PME grid points will be assigned to
>>> vacuum space, for which nothing needs to be done. You can check
>>> imbalances and performance loss in the log file.
>>>
>>> -Justin
>>>
>>>
>>> Thank you
>>>
>>> Sincerely
>>> Sapna
>>>
>>>
>>> -- 
>>>
>>> Justin A. Lemkul
>>> Ph.D. Candidate
>>> ICTAS Doctoral Scholar
>>> MILES-IGERT Trainee
>>> Department of Biochemistry
>>> Virginia Tech
>>> Blacksburg, VA
>>> jalemkul[at]vt.edu  | (540) 231-9080
>>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>>
>>> 
>>> -- gmx-users mailing list gmx-users@gromacs.org
>>> 
>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>> Please search the archive at http://www.gromacs.org/search before
>>> posting!
>>> Please don't post (un)subscribe requests to the list. Use the www
>>> interface or send it to gmx-users-requ...@gromacs.org
>>> .
>>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>>
>>>
>>>
>>>
>>> --
>>> Sapna Sarupria
>>> Post-doctoral Researcher
>>> Princeton University
>>> New Jersey 08540
>>> U.S.A.
>>>
>>> Life isn't about finding yourself. Life is about creating yourself. --
>>> George Bernard Shaw.
>>> Dare to Dream
>>>
>>>
>>
> --
> gmx-users mailing listgmx-users@gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the www interface
> or send it to gmx-users-requ...@gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>



-- 
Sapna Sarupria
Post-doctoral Researcher
Princeton University
New Jersey 08540
U.S.A.

Life isn't about finding yourself. Life is about creating yourself. --
George Bernard Shaw.
Dare to Dream
-- 
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php

Re: [gmx-users] Speeding up simulation

2010-08-18 Thread ms 

On 17/08/10 22:36, Justin A. Lemkul wrote:


There is no way to limit PME to certain ranges. It is a method for
solving infinite sums. Usually PME is a poor choice for in vacuo
simulations and the like, since there's nothing to be done for most of
the simulation box.


This interests me -I've used simple cut-off for in vacuo tests, but I 
still hadn't made my mind on what electrostatic model I should truly 
use. Do you have any suggestion?


thanks,
M.


You may want to see how others (in the literature)
deal with droplet-type simulations.

-Justin



Thanks for your help.

Sapna

On Tue, Aug 17, 2010 at 5:29 PM, Justin A. Lemkul mailto:jalem...@vt.edu>> wrote:



sapna sarupria wrote:

Hi all,

I am running (NVT) simulations of a drop of water (~5 nm in
radius) in a big box (~25 nm) in a box and find that the
simulations are rather slow. I am getting about 0.8 ns per day
when a simulation of bulk system of equivalent number of waters
will be much faster. The number of waters is ~12000. I was
wondering if anyone can suggest methods with which I can speed
up the simulations. I am using domain decomposition and
optimize_fft is set to yes. PME is used for the electrostatics.


I would suspect that the lag comes from a lot of unused PME
calculations. During your run, PME grid points will be assigned to
vacuum space, for which nothing needs to be done. You can check
imbalances and performance loss in the log file.

-Justin


Thank you

Sincerely
Sapna


-- 

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu  | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin


-- gmx-users mailing list gmx-users@gromacs.org

http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before
posting!
Please don't post (un)subscribe requests to the list. Use the www
interface or send it to gmx-users-requ...@gromacs.org
.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php




--
Sapna Sarupria
Post-doctoral Researcher
Princeton University
New Jersey 08540
U.S.A.

Life isn't about finding yourself. Life is about creating yourself. --
George Bernard Shaw.
Dare to Dream





--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/mailing_lists/users.php

Re: [gmx-users] Speeding up simulation

2010-08-17 Thread Gaurav Goel 
Sapna,

Is there a particular reason you're using a box so much larger than the
droplet diameter? If your only concern is that periodic images should not
see each-other maybe you can use a box length=droplet diameter+2*largest
cut-off in the system.

Also using a more spherical box shape such as  rhombic dodecahedron or the
truncated octahedron will help you in reducing the volume of unfilled space.

-Gaurav

On Tue, Aug 17, 2010 at 5:40 PM, sapna sarupria wrote:

> Hi Justin
>
> Thanks for your input. I guess I will do some more literature search and
> try to optimize the simulations as much as I can.
>
> Thanks,
> Sapna
>
>
> On Tue, Aug 17, 2010 at 5:36 PM, Justin A. Lemkul  wrote:
>
>>
>>
>> sapna sarupria wrote:
>>
>>> Hi Justin
>>> Thanks for your response. I suspect that it is the same thing. Is there a
>>> way to make gromacs use PME only within a certain region of the simulation
>>> box? Also, how does one know how to fix the load imbalances (in this
>>> context)?
>>>
>>
>> There is no way to limit PME to certain ranges.  It is a method for
>> solving infinite sums.  Usually PME is a poor choice for in vacuo
>> simulations and the like, since there's nothing to be done for most of the
>> simulation box.  You may want to see how others (in the literature) deal
>> with droplet-type simulations.
>>
>> -Justin
>>
>>
>>> Thanks for your help.
>>>
>>> Sapna
>>>
>>> On Tue, Aug 17, 2010 at 5:29 PM, Justin A. Lemkul >> jalem...@vt.edu>> wrote:
>>>
>>>
>>>
>>>sapna sarupria wrote:
>>>
>>>Hi all,
>>>
>>>I am running (NVT) simulations of a drop of water (~5 nm in
>>>radius) in a big box (~25 nm) in a box and find that the
>>>simulations are rather slow. I am getting about 0.8 ns per day
>>>when a simulation of bulk system of equivalent number of waters
>>>will be much faster. The number of waters is ~12000. I was
>>>wondering if anyone can suggest methods with which I can speed
>>>up the simulations. I am using domain decomposition and
>>>optimize_fft is set to yes. PME is used for the electrostatics.
>>>
>>>
>>>I would suspect that the lag comes from a lot of unused PME
>>>calculations. During your run, PME grid points will be assigned to
>>>vacuum space, for which nothing needs to be done.  You can check
>>>imbalances and performance loss in the log file.
>>>
>>>-Justin
>>>
>>>
>>>Thank you
>>>
>>>Sincerely
>>>Sapna
>>>
>>>
>>>-- 
>>>
>>>Justin A. Lemkul
>>>Ph.D. Candidate
>>>ICTAS Doctoral Scholar
>>>MILES-IGERT Trainee
>>>Department of Biochemistry
>>>Virginia Tech
>>>Blacksburg, VA
>>>jalemkul[at]vt.edu  | (540) 231-9080
>>>
>>>http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>>
>>>
>>>-- gmx-users mailing listgmx-users@gromacs.org
>>>
>>>
>>>http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>Please search the archive at http://www.gromacs.org/search before
>>>posting!
>>>Please don't post (un)subscribe requests to the list. Use the www
>>>interface or send it to gmx-users-requ...@gromacs.org
>>>.
>>>
>>>Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>>
>>>
>>>
>>>
>>> --
>>> Sapna Sarupria
>>> Post-doctoral Researcher
>>> Princeton University
>>> New Jersey 08540
>>> U.S.A.
>>>
>>> Life isn't about finding yourself. Life is about creating yourself. --
>>> George Bernard Shaw.
>>> Dare to Dream
>>>
>>>
>> --
>> 
>>
>> Justin A. Lemkul
>> Ph.D. Candidate
>> ICTAS Doctoral Scholar
>> MILES-IGERT Trainee
>> Department of Biochemistry
>> Virginia Tech
>> Blacksburg, VA
>> jalemkul[at]vt.edu | (540) 231-9080
>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>
>> 
>> --
>> gmx-users mailing listgmx-users@gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at http://www.gromacs.org/search before
>> posting!
>> Please don't post (un)subscribe requests to the list. Use the www
>> interface or send it to gmx-users-requ...@gromacs.org.
>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>
>
>
>
> --
> Sapna Sarupria
> Post-doctoral Researcher
> Princeton University
> New Jersey 08540
> U.S.A.
>
> Life isn't about finding yourself. Life is about creating yourself. --
> George Bernard Shaw.
> Dare to Dream
>
> --
> gmx-users mailing listgmx-users@gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-requ...@gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lis

Re: [gmx-users] Speeding up simulation

2010-08-17 Thread sapna sarupria 
Hi Justin

Thanks for your input. I guess I will do some more literature search and try
to optimize the simulations as much as I can.

Thanks,
Sapna

On Tue, Aug 17, 2010 at 5:36 PM, Justin A. Lemkul  wrote:

>
>
> sapna sarupria wrote:
>
>> Hi Justin
>> Thanks for your response. I suspect that it is the same thing. Is there a
>> way to make gromacs use PME only within a certain region of the simulation
>> box? Also, how does one know how to fix the load imbalances (in this
>> context)?
>>
>
> There is no way to limit PME to certain ranges.  It is a method for solving
> infinite sums.  Usually PME is a poor choice for in vacuo simulations and
> the like, since there's nothing to be done for most of the simulation box.
>  You may want to see how others (in the literature) deal with droplet-type
> simulations.
>
> -Justin
>
>
>> Thanks for your help.
>>
>> Sapna
>>
>> On Tue, Aug 17, 2010 at 5:29 PM, Justin A. Lemkul > jalem...@vt.edu>> wrote:
>>
>>
>>
>>sapna sarupria wrote:
>>
>>Hi all,
>>
>>I am running (NVT) simulations of a drop of water (~5 nm in
>>radius) in a big box (~25 nm) in a box and find that the
>>simulations are rather slow. I am getting about 0.8 ns per day
>>when a simulation of bulk system of equivalent number of waters
>>will be much faster. The number of waters is ~12000. I was
>>wondering if anyone can suggest methods with which I can speed
>>up the simulations. I am using domain decomposition and
>>optimize_fft is set to yes. PME is used for the electrostatics.
>>
>>
>>I would suspect that the lag comes from a lot of unused PME
>>calculations. During your run, PME grid points will be assigned to
>>vacuum space, for which nothing needs to be done.  You can check
>>imbalances and performance loss in the log file.
>>
>>-Justin
>>
>>
>>Thank you
>>
>>Sincerely
>>Sapna
>>
>>
>>-- 
>>
>>Justin A. Lemkul
>>Ph.D. Candidate
>>ICTAS Doctoral Scholar
>>MILES-IGERT Trainee
>>Department of Biochemistry
>>Virginia Tech
>>Blacksburg, VA
>>jalemkul[at]vt.edu  | (540) 231-9080
>>
>>http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>
>>
>>-- gmx-users mailing listgmx-users@gromacs.org
>>
>>
>>http://lists.gromacs.org/mailman/listinfo/gmx-users
>>Please search the archive at http://www.gromacs.org/search before
>>posting!
>>Please don't post (un)subscribe requests to the list. Use the www
>>interface or send it to gmx-users-requ...@gromacs.org
>>.
>>
>>Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>
>>
>>
>>
>> --
>> Sapna Sarupria
>> Post-doctoral Researcher
>> Princeton University
>> New Jersey 08540
>> U.S.A.
>>
>> Life isn't about finding yourself. Life is about creating yourself. --
>> George Bernard Shaw.
>> Dare to Dream
>>
>>
> --
> 
>
> Justin A. Lemkul
> Ph.D. Candidate
> ICTAS Doctoral Scholar
> MILES-IGERT Trainee
> Department of Biochemistry
> Virginia Tech
> Blacksburg, VA
> jalemkul[at]vt.edu | (540) 231-9080
> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>
> 
> --
> gmx-users mailing listgmx-users@gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the www interface
> or send it to gmx-users-requ...@gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>



-- 
Sapna Sarupria
Post-doctoral Researcher
Princeton University
New Jersey 08540
U.S.A.

Life isn't about finding yourself. Life is about creating yourself. --
George Bernard Shaw.
Dare to Dream
-- 
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php

Re: [gmx-users] Speeding up simulation

2010-08-17 Thread Justin A. Lemkul 



sapna sarupria wrote:

Hi Justin
Thanks for your response. I suspect that it is the same thing. Is there 
a way to make gromacs use PME only within a certain region of the 
simulation box? Also, how does one know how to fix the load imbalances 
(in this context)?


There is no way to limit PME to certain ranges.  It is a method for solving 
infinite sums.  Usually PME is a poor choice for in vacuo simulations and the 
like, since there's nothing to be done for most of the simulation box.  You may 
want to see how others (in the literature) deal with droplet-type simulations.


-Justin



Thanks for your help.

Sapna

On Tue, Aug 17, 2010 at 5:29 PM, Justin A. Lemkul > wrote:




sapna sarupria wrote:

Hi all,

I am running (NVT) simulations of a drop of water (~5 nm in
radius) in a big box (~25 nm) in a box and find that the
simulations are rather slow. I am getting about 0.8 ns per day
when a simulation of bulk system of equivalent number of waters
will be much faster. The number of waters is ~12000. I was
wondering if anyone can suggest methods with which I can speed
up the simulations. I am using domain decomposition and
optimize_fft is set to yes. PME is used for the electrostatics.


I would suspect that the lag comes from a lot of unused PME
calculations. During your run, PME grid points will be assigned to
vacuum space, for which nothing needs to be done.  You can check
imbalances and performance loss in the log file.

-Justin


Thank you

Sincerely
Sapna


-- 



Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu  | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin


-- 
gmx-users mailing listgmx-users@gromacs.org


http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before
posting!
Please don't post (un)subscribe requests to the list. Use the www
interface or send it to gmx-users-requ...@gromacs.org
.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php




--
Sapna Sarupria
Post-doctoral Researcher
Princeton University
New Jersey 08540
U.S.A.

Life isn't about finding yourself. Life is about creating yourself. -- 
George Bernard Shaw.

Dare to Dream



--


Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin


--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/mailing_lists/users.php

Re: [gmx-users] Speeding up simulation

2010-08-17 Thread sapna sarupria 
Hi Justin
Thanks for your response. I suspect that it is the same thing. Is there a
way to make gromacs use PME only within a certain region of the simulation
box? Also, how does one know how to fix the load imbalances (in this
context)?

Thanks for your help.

Sapna

On Tue, Aug 17, 2010 at 5:29 PM, Justin A. Lemkul  wrote:

>
>
> sapna sarupria wrote:
>
>> Hi all,
>>
>> I am running (NVT) simulations of a drop of water (~5 nm in radius) in a
>> big box (~25 nm) in a box and find that the simulations are rather slow. I
>> am getting about 0.8 ns per day when a simulation of bulk system of
>> equivalent number of waters will be much faster. The number of waters is
>> ~12000. I was wondering if anyone can suggest methods with which I can speed
>> up the simulations. I am using domain decomposition and optimize_fft is set
>> to yes. PME is used for the electrostatics.
>>
>
> I would suspect that the lag comes from a lot of unused PME calculations.
> During your run, PME grid points will be assigned to vacuum space, for which
> nothing needs to be done.  You can check imbalances and performance loss in
> the log file.
>
> -Justin
>
>
>> Thank you
>>
>> Sincerely
>> Sapna
>>
>>
> --
> 
>
> Justin A. Lemkul
> Ph.D. Candidate
> ICTAS Doctoral Scholar
> MILES-IGERT Trainee
> Department of Biochemistry
> Virginia Tech
> Blacksburg, VA
> jalemkul[at]vt.edu | (540) 231-9080
> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>
> 
> --
> gmx-users mailing listgmx-users@gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the www interface
> or send it to gmx-users-requ...@gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>



-- 
Sapna Sarupria
Post-doctoral Researcher
Princeton University
New Jersey 08540
U.S.A.

Life isn't about finding yourself. Life is about creating yourself. --
George Bernard Shaw.
Dare to Dream
-- 
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php

Re: [gmx-users] Speeding up simulation

2010-08-17 Thread Justin A. Lemkul 



sapna sarupria wrote:

Hi all,

I am running (NVT) simulations of a drop of water (~5 nm in radius) in a 
big box (~25 nm) in a box and find that the simulations are rather slow. 
I am getting about 0.8 ns per day when a simulation of bulk system of 
equivalent number of waters will be much faster. The number of waters is 
~12000. I was wondering if anyone can suggest methods with which I can 
speed up the simulations. I am using domain decomposition and 
optimize_fft is set to yes. PME is used for the electrostatics.


I would suspect that the lag comes from a lot of unused PME calculations. 
During your run, PME grid points will be assigned to vacuum space, for which 
nothing needs to be done.  You can check imbalances and performance loss in the 
log file.


-Justin



Thank you

Sincerely
Sapna



--


Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin


--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/mailing_lists/users.php