Re: increasing MAX_EXCLUSIONS

From: Francesco Oteri (francesco.oteri_at_gmail.com)
Date: Thu Jan 20 2011 - 09:56:35 CST

I solved the problem creatin a new topology.
In particular the problematic topology contains the DNA and then the
protein.
The new topology contains at the begin the protein and then the DNA.
Swapping the position of protein and DNA solved the problem

Il 19/01/2011 23:07, Francesco Oteri ha scritto:
> Dear namd users,
> I need a suggestion regarding increasing of MAX_EXCLUSIONS variable.
>
> In particular, I need to simulate a system composed by about 1.100.000
> atoms.
> When I try to minimize such a system with namd-gpu, I obtain the
> famous error:
>
> "........Increase MAX_EXCLUSIONS."
>
>
> In the file ComputeNonbondedCUDAKernel.h, the value of
> MAX_EXCLUSIONS is 2048. I increased this value to 6144 (1024*6) and
> the compilation has been succesfully completed, but when I run a
> minimization with constrain on Protein-DNA atoms,
> NAMD destroys the protein increasing the lenght of the covalent bonds.
> The CPU version doesn't show this behaviour.
>
> I think that simply increasing MAX_EXCLUSIONS is not enough :)
>
> Does anyone have a solution to the problem?
>
> Thanks,
>
>
>
> Francesco
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 05:23:31 CST