[Pw_forum] A few questions about MD simulations

[Mike Marchywka]

----------------------------------------
> To: pw_forum at pwscf.org
> From: akohlmey at gmail.com
> Date: Sat, 1 Jun 2013 17:50:07 +0000
> Subject: Re: [Pw_forum] A few questions about MD simulations
>
> Nevertheless significant compression should be possible using some knowledge of the physics, e.g. through projecting on atomic wavefunctions. Mind you, this is a lossy compression if a minimal basis is used. I've seen pretty good compression and a decent completeness for example by projecting on an STO-6G basis for example. That was with a different coff though, but still using plane waves and DFT.

Well, the wave functions are hardly random or uniformly distributed over the full range of numeric values and simple things like smoothness can be exploited for somec ompression that is  easy and
fairly fast. Getting exact results from floating point algorithms can be difficult but if you can tolerate a few
changed LSB's or want to put some thought into it it may be something to consider :)



>
> Axel
> ------Original Message------
> From: Paolo Giannozzi
> Sender: pw_forum-bounces at pwscf.org
> To: PWSCF Forum
> ReplyTo: PWSCF Forum
> Subject: Re: [Pw_forum] A few questions about MD simulations
> Sent: Jun 1, 2013 7:36 PM
>
> On Sat, 2013-06-01 at 10:06 -0400, Mike Marchywka wrote:
>
>> Don't ignore the possibility of adding compression code,
>> something like gzip
>
> large data files are written in binary format, so there
> is little to be gained from compression.
>
> P.
> --
> Paolo Giannozzi, Dept. Chemistry&Physics&Environment,
> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
> Phone +39-0432-558216, fax +39-0432-558222
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>
>
> --
> Dr. Axel Kohlmeyer akohlmey at gmail.com http://goo.gl/1wk0
> International Centre for Theoretical Physics, Trieste. Italy.
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum