Re: cgmd stability

From: BIN ZHANG (zhngbn_at_gmail.com)
Date: Thu Feb 19 2009 - 00:31:35 CST

The protein does deform, but not going wild.
>
> Please keep in mind that people haven't really done a whole
> lot of SMD work using this model.

Why would this matter? I thought SMD should have nothing to do with
the model I use (whether all-atom or coarse-grained).

> Are you pulling by one bead or the
> center of mass of several? Also, do the crashes happen consistently
> around the same time if you go repeat the crashed run?
>
What do you mean by repeat? I did several simulation from the same
configuration, but with different random numbers, and they do NOT
crash at the same time. But I didn't really test "same configuration,
same random number".

Thanks again.

Bin

>
> BIN ZHANG wrote:
>> Hi Peter:
>>
>> Thanks for your reply.
>> I'm using 1-2 exclusions, and the dynamics is indeed stable if I have
>> no pulling at all. You are also right in the fact that the stability
>> depends on the pulling velocity. However, it seems to be more stable
>> when I pull it faster, which does NOT really make that much sense
>> to me.
>> Can you be more specific about "deforming"? unphysically long bonds?
>>
>> Best,
>> Bin
>>
>>
>> On Feb 18, 2009, at 6:05 PM, Peter Freddolino wrote:
>>
>>> Hi Bin,
>>> can I verify that you're using 1-2 exclusions and not 1-4? RBCG is
>>> designed to go with 1-2 exclusions, and any other setting can
>>> certainly
>>> cause the problems that you're seeing. If that is not the problem,
>>> I'd
>>> recommend looking at your system immediately before you get
>>> crashes and
>>> seeing if things are becoming visibly deformed. Have you tried
>>> running
>>> without the applied pulling, or with a slower pulling velocity?
>>>
>>> Peter
>>>
>>> BIN ZHANG wrote:
>>>> Dear all:
>>>>
>>>> I've been experiencing with rbcg using NAMD for a while, and
>>>> often got
>>>> the error "FATAL ERROR: Bad global exclusion count!" during the
>>>> simulation, even with a time step as small as 5 fs. (This could
>>>> also
>>>> partially blame the calculation I was doing probably, pulling
>>>> peptide
>>>> through membrane)
>>>>
>>>> Searching about the list, I found the following reply from peter:
>>>>
>>>>> Hi Jiaqi,
>>>>> it is important to understand the source of the bad global
>>>>> exclusion
>>>>> count error: Two atoms which should have an exclusion are not on
>>>>> neighboring patches. This happens a lot in coarse grained systems
>>>>> because they have such long bonds. What do you have exclude set
>>>>> to?
>>>>> And
>>>>> how far apart are your particles?
>>>>> Peter
>>>>
>>>> If this is the case, I wonder would decrease the number of cpus
>>>> used,
>>>> or increase the "margin" parameter help in solving the problem? Is
>>>> there other way to systemically reduce the error occurring?
>>>>
>>>> BTW, I do NOT think this error occurs due to bad initial
>>>> configuration, because it could happen after several ten ns
>>>> running.
>>>>
>>>> Thanks a lot.
>>>> Bin
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> -------------------------------------------------------------
>> The tree of liberty must be refreshed from time to time with the
>> blood
>> of patriots and tyrants.

-------------------------------------------------------------
The tree of liberty must be refreshed from time to time with the blood
of patriots and tyrants.

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:50:31 CST