From: Joshua D. Moore (jdmoore_at_unity.ncsu.edu)
Date: Wed Mar 29 2006 - 15:37:24 CST
In my experience I wasn't able to recover the files. It was just a wasted
run (and a rather large one :-P).
Lea, in my experience, the 64-bit CatDCD only worked with DCD files
written with 64-bit NAMD and the 32-bit CatDCD only worked with DCD files
written with 32-bit NAMD. So if you want to try to split your files, you
probably would have to use a 32-bit version of CatDCD since your DCD files
were written with the 32-bit version of NAMD.
I don't begin to understand the details of all of this, but this is just
what I found from trial and error.
Josh
On Wed, March 29, 2006 12:55 pm, jonathan_at_ibt.unam.mx wrote:
> Try opening it with a hexadecimal editor ("hexedit" in Linux), see if you
> can
> read the title or, if you have some understanding of the format, try
> extracting
> a few coordinates at the end by hand and see if they make sense.
>
> Any way, I guess the best choice is to split the original file in the
> cluster
> (or wherever the simulation was run) into less than 2GB pieces, and then
> transfer them to your workstation.
>
> J. Valencia
>
> Quoting Lea Thgersen <lea_at_birc.au.dk>:
>
>> Can it really be that I have lost everything because the file got larger
>> than 2Gb???
>>
>> I cannot read in any way the DCD file from a 2ns simulation executed
>> with the precompiled binary called NAMD_2.6b1_AIX-POWER-MPI on an IBM
>> rs6000 pSeries 690 with AIX5.1D OS and tcl8.3.
>>
>> During the simulation I have frequently copied the dcd-file to my
>> workstation (AMD64), used flipdcd, and studied the progress. However,
>> since the size of the dcd-file got lager than 2Gb I haven't been able to
>> open it in any way. Of course every time I copied the dcd-file to my
>> workstation, the older version was overwritten (stupid, stupid me) so
>> when the file suddenly gave me problems, I had no backup.
>>
>> When trying to use flipdcd on my workstation it gave the following:
>> >flipdcd dmpc_sol2.eq.dcd
>> dmpc_sol2.eq.dcd is not in DCD format
>>
>> I can understand it will give problems since I only have 2Gb memory on
>> my workstation and flipdcd uses an amount of memory corresponding to the
>> filesize (as far as I understand).
>>
>> When trying to use catdcd on my workstation it gave the following:
>> >catdcd dmpc_sol2.eq.dcd
>> CatDCD 4.0
>> read_dcdheader returned -6
>> Error: could not open file 'dmpc_sol2.eq.dcd' for reading.
>>
>> I guessed that maybe catdcd need it to be in native endian? Anyhow, I
>> found some previous post on this list about problems with large dcd
>> files from AIX:
>> http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l/3332.html
>>
>> Thereby inspired I got the latest cvs version of the plugin-tree for vmd
>> and compiled a 64bit version of catdcd on the machine where the
>> simulation was executed.
>> However, whether I compile catdcd with gmake AIX5 or AIX5_64, I get the
>> following output when trying to use it on my file:
>>
>> >catdcd dmpc_sol2.eq.dcd
>> CatDCD 4.0
>> dcdplugin) detected standard 32-bit DCD file of native endianness
>> dcdplugin) read_dcdheader: corruption or unrecognized file structure
>> Error: could not open file 'dmpc_sol2.eq.dcd' for reading.
>>
>> When trying to use flipdcd on the AIX machine I get the following
>> >flipdcd dmpc_sol2.eq.dcd
>> Can't stat dmpc_sol2.eq.dcd
>>
>>
>> So again I ask you: Can it really be that I have lost everything because
>> the file got larger than 2Gb???
>>
>> Thank you for any help
>> Lea
>>
>
>
>
>
> ----------------------------------------------------------------
> Este mensaje fue enviado desde el servidor Webmail del Instituto de
> Biotecnologia.
>
--------------------------------------------------
Joshua D. Moore
Graduate Student
North Carolina State University
Department of Chemical and Biomolecular Engineering
Box 7905 Centennial Campus
Engineering Building I
911 Partners Way
Raleigh, NC 27695-7905
Phone: (919) 513-2051
Fax: (919) 513-2470
Email: jdmoore_at_unity.ncsu.edu
--------------------------------------------------
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:43:27 CST