<div>From Anton's advices, huge force comes from a small z (with relatively large x and y), but real r is larger than z which should give a normal force. (If I don't understand wrong, those particles are exposed to unreasonable close interactions)
<br><br>Talking about the real r2<br>If z < cutoff but r2 > cutoff, it's not in the neighborlist (a)<br>If z > cutoff, it's anyway not in the neighborlist (b)<br>only real r2<cutoff will appear in neighborlist. (c)
<br><br>Currently, only type (c) is in the neighborlist. Some border case of type (a) (relatively large x,y enter the neighborlist) seems to cause problems. So might be a good trial to eliminate those?<br><br> </div>
<div><span class="gmail_quote">On 10/18/06, <b class="gmail_sendername">David van der Spoel</b> <<a href="mailto:spoel@xray.bmc.uu.se">spoel@xray.bmc.uu.se</a>> wrote:</span>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">Chenyue Xing wrote:<br>> Yeah. I printed r^2 out as well as all Zs. r^2 is really calculated<br>> based on Z.
<br><br>i am talking about the real r2, for those particles not in the<br>neighborlist. you have to modify neigborsearching as well.<br><br>please reread carefully Antons comments.<br><br>> So there's a r^2 calculation in neighborsearching too and it's somehow
<br>> used in table look-up?<br>><br>><br>> On 10/18/06, *David van der Spoel* <<a href="mailto:spoel@xray.bmc.uu.se">spoel@xray.bmc.uu.se</a><br>> <mailto:<a href="mailto:spoel@xray.bmc.uu.se">spoel@xray.bmc.uu.se
</a>>> wrote:<br>><br>> Chenyue Xing wrote:<br>> > Hi Anton,<br>> ><br>> > Thanks for the nice suggestions.<br>> > As for the first possibility, I've been investigating and printed
<br>> out<br>> > all the zi, zj, but haven't got a clue yet.<br>> > As for the 2nd, the modified code still works in the same manner,<br>> that<br>> > rsq = z^2 and then take the sqrt, so a negative r should not be a
<br>> problem.<br>> ><br>><br>><br>> have you checked r2 as well? I'm pretty sure it is the<br>> nieghboursearching.<br>><br>> > If you have any other guess, please let me know. That will help a
<br>> lot.<br>> ><br>> > Thank you !<br>> > Chenyue<br>> ><br>> ><br>> > On 10/18/06, *Anton Feenstra* <<a href="mailto:feenstra@few.vu.nl">feenstra@few.vu.nl
</a><br>> <mailto:<a href="mailto:feenstra@few.vu.nl">feenstra@few.vu.nl</a>><br>> > <mailto:<a href="mailto:feenstra@few.vu.nl">feenstra@few.vu.nl</a> <mailto:<a href="mailto:feenstra@few.vu.nl">
feenstra@few.vu.nl</a>>>> wrote:<br>> ><br>> > Chenyue Xing wrote:<br>> > [...]<br>> > > Thank you for your discussion and I hope I can provide<br>> more info
<br>> > soon so<br>> > > that you might be able to give me more advices.<br>> ><br>> > There are two problems I can think of that may be bothering you.<br>> > First is related to what David already mentioned: if you have
<br>> an LJ-type<br>> > interaction and your z becomes very small, the resulting<br>> force is HUGE.<br>> > This can happen suddenly in your case when a particle enters<br>> the cutoff
<br>> > (as determined by the neighbourlist update), with relatively<br>> large x or<br>> > y, but small z.<br>> > Second is that r=z is not really equivalent to the 3D case,
<br>> because<br>> > your<br>> > r can become negative (I haven't looked at the code, but I<br>> suspect<br>> > something like z=zi-zj for particle i and j). In the 3D case,
<br>> you have<br>> > the squares and square-root. I can imagine that a negative r<br>> will not be<br>> > present in the lookup table ;-)<br>> ><br>> ><br>
> > --<br>> > Groetjes,<br>> ><br>> > Anton<br>> ><br>> > _____________<br>> _______________________________________________________
<br>> > |<br>> | |<br>> > | _ _ ___,| K. Anton<br>> Feenstra |<br>> > | / \ / \'| | | IBIVU/Bioinformatics - Vrije Universiteit
<br>> Amsterdam |<br>> > |( | )| | | De Boelelaan 1083 - 1081A HV Amsterdam -<br>> Netherlands |<br>> > | \_/ \_/ | | | Tel +31 20 59 87783 - Fax +31 20 59 87653 -<br>> Room P440 |
<br>> > | | <a href="mailto:Feenstra@few.vu.nl">Feenstra@few.vu.nl</a><br>> <mailto:<a href="mailto:Feenstra@few.vu.nl">Feenstra@few.vu.nl</a>> <mailto:<a href="mailto:Feenstra@few.vu.nl">
Feenstra@few.vu.nl</a><br>> <mailto:<a href="mailto:Feenstra@few.vu.nl">Feenstra@few.vu.nl</a>>> -<br>> > <a href="http://www.few.vu.nl/~feenstra/">www.few.vu.nl/~feenstra/</a> <<a href="http://www.few.vu.nl/~feenstra/">
http://www.few.vu.nl/~feenstra/</a>><br>> <<a href="http://www.few.vu.nl/~feenstra/">http://www.few.vu.nl/~feenstra/</a>> |<br>> > | | "If You See Me Getting High, Knock Me Down"
<br>> (RHCP) |<br>> ><br>> |_____________|_______________________________________________________|<br>> > _______________________________________________<br>> > gmx-developers mailing list
<br>> > <a href="mailto:gmx-developers@gromacs.org">gmx-developers@gromacs.org</a><br>> <mailto:<a href="mailto:gmx-developers@gromacs.org">gmx-developers@gromacs.org</a>><br>> <mailto:
<a href="mailto:gmx-developers@gromacs.org">gmx-developers@gromacs.org</a> <mailto:<a href="mailto:gmx-developers@gromacs.org">gmx-developers@gromacs.org</a>>><br>> > <a href="http://www.gromacs.org/mailman/listinfo/gmx-developers">
http://www.gromacs.org/mailman/listinfo/gmx-developers</a><br>> <<a href="http://www.gromacs.org/mailman/listinfo/gmx-developers">http://www.gromacs.org/mailman/listinfo/gmx-developers</a>><br>> > <
<a href="http://www.gromacs.org/mailman/listinfo/gmx-developers">http://www.gromacs.org/mailman/listinfo/gmx-developers</a>><br>> > Please don't post (un)subscribe requests to the list. Use the<br>> > www interface or send it to
<br>> <a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org</a><br>> <mailto:<a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org</a>>
<br>> > <mailto:<a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org</a><br>> <mailto:<a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org
</a>>>.<br>> ><br>> ><br>> ><br>> ><br>> ------------------------------------------------------------------------<br>> ><br>> > _______________________________________________
<br>> > gmx-developers mailing list<br>> > <a href="mailto:gmx-developers@gromacs.org">gmx-developers@gromacs.org</a> <mailto:<a href="mailto:gmx-developers@gromacs.org">gmx-developers@gromacs.org
</a>><br>> > <a href="http://www.gromacs.org/mailman/listinfo/gmx-developers">http://www.gromacs.org/mailman/listinfo/gmx-developers</a><br>> > Please don't post (un)subscribe requests to the list. Use the
<br>> > www interface or send it to <a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org</a><br>> <mailto:<a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org
</a>>.<br>><br>><br>> --<br>> David.<br>> ________________________________________________________________________<br>><br>> David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
<br>> Dept. of Cell and Molecular Biology, Uppsala University.<br>> Husargatan 3, Box 596, 75124 Uppsala, Sweden<br>> phone: 46 18 471 4205 fax: 46 18 511 755<br>> <a href="mailto:spoel@xray.bmc.uu.se">
spoel@xray.bmc.uu.se</a><br>> <mailto:<a href="mailto:spoel@xray.bmc.uu.se">spoel@xray.bmc.uu.se</a>> <a href="mailto:spoel@gromacs.org">spoel@gromacs.org</a><br>> <mailto:<a href="mailto:spoel@gromacs.org">
spoel@gromacs.org</a>> <a href="http://folding.bmc.uu.se">http://folding.bmc.uu.se</a><br>> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++<br>><br>> _______________________________________________
<br>> gmx-developers mailing list<br>> <a href="mailto:gmx-developers@gromacs.org">gmx-developers@gromacs.org</a> <mailto:<a href="mailto:gmx-developers@gromacs.org">gmx-developers@gromacs.org</a>><br>
> <a href="http://www.gromacs.org/mailman/listinfo/gmx-developers">http://www.gromacs.org/mailman/listinfo/gmx-developers</a><br>> Please don't post (un)subscribe requests to the list. Use the<br>> www interface or send it to
<a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org</a><br>> <mailto:<a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org</a>>.<br>><br>
><br>><br>> ------------------------------------------------------------------------<br>><br>> _______________________________________________<br>> gmx-developers mailing list<br>> <a href="mailto:gmx-developers@gromacs.org">
gmx-developers@gromacs.org</a><br>> <a href="http://www.gromacs.org/mailman/listinfo/gmx-developers">http://www.gromacs.org/mailman/listinfo/gmx-developers</a><br>> Please don't post (un)subscribe requests to the list. Use the
<br>> www interface or send it to <a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org</a>.<br><br><br>--<br>David.<br>________________________________________________________________________
<br>David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,<br>Dept. of Cell and Molecular Biology, Uppsala University.<br>Husargatan 3, Box 596, 75124 Uppsala, Sweden<br>phone: 46 18 471 4205 fax: 46 18 511 755
<br><a href="mailto:spoel@xray.bmc.uu.se">spoel@xray.bmc.uu.se</a> <a href="mailto:spoel@gromacs.org">spoel@gromacs.org</a> <a href="http://folding.bmc.uu.se">http://folding.bmc.uu.se</a><br>++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
<br>_______________________________________________<br>gmx-developers mailing list<br><a href="mailto:gmx-developers@gromacs.org">gmx-developers@gromacs.org</a><br><a href="http://www.gromacs.org/mailman/listinfo/gmx-developers">
http://www.gromacs.org/mailman/listinfo/gmx-developers</a><br>Please don't post (un)subscribe requests to the list. Use the<br>www interface or send it to <a href="mailto:gmx-developers-request@gromacs.org">gmx-developers-request@gromacs.org
</a>.<br></blockquote></div><br>