[gmx-developers] inner loop and lennard-jones

[Erik Lindahl]

Hi Daniele,

First, check the flopcount statistics at the end of your run for a  
summary of which innerloops were used. Then check  
gromacs/include/types/nrnb.h for an explanation of the innerloop  
enumeration.

The 1,4-interactions are always tabulated for performance reasons  
(inl3300), but that loop might not be used for the normal nonbonded  
interactions.

You should also be aware that we use assembly loops on x86 hardware;  
you can turn them off by setting the variable NOASSEMBLYLOOPS before  
starting mdrun.

Cheers,

Erik


On Wednesday, Apr 30, 2003, at 10:37 US/Pacific, Daniele Bemporad wrote:

> Hi all.
> I need to hack the gromacs code and modify the lennard-jones  
> potential. I
> have added print statements in each inner loop to see which of them was
> involved in my simulation and to check the value of relevant variables.
> In particular, it has turned out that the inner loop involved is the
> inl3300. However, only the 1-4 interactions are printed out and there  
> is
> no way to find where and when the other non-bonded interactions are
> calculated. Can anybody be of any help?
> Thanks a lot
> Daniele Bemporad
> _______________________________________________
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-developers
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-developers-request at gromacs.org.
>
------------------------------------------------------------------------ 
-----
Erik Lindahl, MSc, PhD     <lindahl at stanford.edu>
D109, Fairchild Building
Dept. Structural Biology, Stanford University School of Medicine
Tel. 650-7250754    Fax. 650-7238464