Dear Justin thanks for your answer
actually I'm trying to reproduce other constant surface tension simulations
of bilayers to make sure my method is right. they calculate area per lipid
for different surface tensions from approximately 0 to 100 dyn/cm and that
is why I've mdp files with different
I ran equilibrium simulations, and then a longer simulations with the
recommended sd integrator. Still, the mass_lambda dH/dL is nonzero and
sometimes it is the dominant contribution at certain lambdas. I don't think
it is a convergence issue because there aren't huge fluctuations in dH/dL
values
On 5/12/17 12:27 PM, Ali Shomali wrote:
Dear Justin
Is there any possibility that the cause of problem is my simulation time ?
(3 ns)
Yes, that's about 50 times too short to be credible. Lipids relax over tens of
nanoseconds, and typical production run times for well equilibrated membrane
Dear Justin
Is there any possibility that the cause of problem is my simulation time ?
(3 ns)
On Mon, May 8, 2017 at 4:38 PM, Justin Lemkul wrote:
>
>
> On 5/7/17 2:56 PM, Ali Shomali wrote:
>
>> Thanks so much Justin for your answer. actually I performed different
>>
Dear Justin thanks for the answer
I think the surface tension line was deleted in email by mistake , sorry
for that. I managed to perform other simulations for DMPC and DPPC . in all
of them I applied reported surface tensions ( twice the reported surface
tension per interface) in all of them my
Hi,
There already is an applicable Redmine (
https://redmine.gromacs.org/issues/2061). Ake's observations agreed with
two CUDA 8.0 installs I checked, so it looks like CUDA 8.0 acquired the
header and stub library, but the driver package still has the real library.
We can't really / shouldn't try
Also, knowing the group you selected in the mdp for writing the xtc file is
key to understanding why what you get might not match what you are trying
to do...
Mark
On Fri, May 12, 2017 at 1:51 PM Justin Lemkul wrote:
>
>
> On 5/11/17 1:43 PM, Franco Henrique wrote:
> > Dear
On 5/11/17 1:43 PM, Franco Henrique wrote:
Dear Justin Lemkul,
Thank’s for prompt reply.
My protein has 198 aminoacides (PDB: 2GOJ;
http://www.rcsb.org/pdb/explore.do?structureId=2GOJ). After RMSF command (gmx
rmsf -f .xtc -s .tpr -o .xvg -res), the output has only 103 aminoacides as
CUDA 8.0.61_375_26 does not contain the actual NVML lib, only stubs.
And NVML is still separate from the CUDA package as far as i can see.
The libnvidia-ml.so in /lib64/stubs is much smaller than the
actual library.
On 05/12/2017 01:11 PM, Szilárd Páll wrote:
> Mat,
>
> Thanks for the report. I
Mat,
Thanks for the report. I was under the impression that this should have
"just worked" given that NVML is now in the toolkit, but if it does not, we
should look into it.
Can you please, file an issue on redmine.gromacs.org (with reproducing
steps).
Cheers,
--
Szilárd
On Fri, May 12, 2017
CUDA starting from the version 8.0 is not separated from the NVML library
any more, and cannot be downloaded independently. NVML is part of the
toolkit. However, the documentation requires the
GPU_DEPLOYMENT_KIT_ROOT_DIR to find NVML.
I have tried setting the variable to multiple combinations
Mark,
Thanks Mark! Indeed it was the problem. I forgot to update residuetypes.dat.
Regards,
Simon
2017-05-12 15:46 GMT+08:00 Mark Abraham :
> Hi,
>
> Those define your residue, but do not classify it, so pdb2gmx doesn't know
> that protein-specific termini are
Hi,
Those define your residue, but do not classify it, so pdb2gmx doesn't know
that protein-specific termini are appropriate.
>From the link I provided: "Add your residue to residuetypes.dat with the
appropriate specification (Protein, DNA, Ion, etc)."
Mark
On Fri, May 12, 2017 at 9:39 AM
Dear Mark,
Thanks for your prompt reply. I have already updated my .rtp and .hdb for
my new residue. Would you mind to suggest which file I am missing?
I am happy to provide relevant parts of .rtp and .hdb if necessary.
Regards,
Simon
2017-05-12 15:31 GMT+08:00 Mark Abraham
Hi,
I suspect you need to add your residue to the class of Protein residues,
per
http://www.gromacs.org/Documentation/How-tos/Adding_a_Residue_to_a_Force_Field
Mark
On Fri, May 12, 2017 at 9:05 AM Simon Kit Sang Chu
wrote:
> Hi everyone,
>
> I am using a new forcefield
Hi everyone,
I am using a new forcefield which does not give proper neutral capping, but
requires me to create a custom made residue for that purpose. Specifically,
I want a NH2 cap instead of NH+.
I rename VAL1 into VAT1, update .hdb & .rtp. However, when I pdb2gmx -ter
to create a topology
16 matches
Mail list logo