AMBER Archive (2006)

Subject: RE: AMBER: GBSA EGB = NaN

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Thu Aug 17 2006 - 21:57:29 CDT

Dear William,

Can you post the input files you used both for equilibration and the GB part
of the run. Also include the first few steps of the output file so we can
look more carefully. What happens if you run the calculation from the
beginning with GB turned on? E.g. during the initial minimisation - do you
get a problem here? Also do all the GB test cases work? And how big is your
molecule? What is you hardware (compilers, math library etc). I have come
across bugs in IBM's MASSV library on Blue Gene systems where you would get
NaN for GB simulations on more than about 20,000 atoms (vexp on vector
lengths of > 10^8 gives NaN's) but this is a very specific case. There could
be other systems that have bugs in their vector math libraries though. But
first of all before we jump to conclusions it would be good to see all your
input and output files.

All the best
Ross

/\
\/
|\oss Walker

| HPC Consultant and Staff Scientist |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

> -----Original Message-----
> From: owner-amber_at_scripps.edu
> [mailto:owner-amber_at_scripps.edu] On Behalf Of William Lianhu Wei
> Sent: Thursday, August 17, 2006 19:31
> To: amber_at_scripps.edu
> Subject: AMBER: GBSA EGB = NaN
>
> Dear all,
>
> I finished molecular dynamics simulation of my system about
> 3ns. From the
> energy profile, the system reached equilibrium at about 1.5
> ns. I was trying
> to calculate the binding free energy. The trajectories I
> used was after
> 1.5ns. No matter which trajectory I used, I always got
> EGB=NaN (on Complex
> and receptor), and the program stopped at statistical
> calculation. The VDW
> contribution is okay. Does anyone meet this before, or give
> me any clue for
> that. In 2004 archive, there is somebody asked this, but
> there was no reply
> for this. Thanks in advance.
>
> William
>
> --
> --------------------------------------------------------------
> William Wei Tel: +1-416-946-7551
> Faculty of Pharmacy Email: william.wei_at_utoronto.ca
> University of Toronto weilianhu_at_hotmail.com
> 19 Russell St.
> Toronto, Ontario
> M5S 2S2,Canada
> --------------------------------------------------------------
> ---------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu