Today : Fri, Jul 21 2017 .


Fast Multipole Methods (CTP)


Program - Article difference


Page 1

Posts 4




Post: 001  

by  Lityk    on 20.12.2011 - 21:41

1. Thank you very much for the implementation !!!
2. In Article Authors use
Exp(-beta R)/ R
However in the program they left beta also near R in denominator:
Exp(-beta R)/ (beta R)

Probably that should be corrected in the program.confused smiley

Part of program:

RDIS = DSQRT(RR)*BETA
EXPR= DEXP(-RDIS)
TERM1 = CHARGE(JAT)/RDIS*EXPR*PI/2.0D0

Profile    

Ahha245

Post: 002  

23.02.2012 - 02:37

hi Lityk

could you tell me the compiler that you have used to run the code? i've run it on Ubuntu but not successful :(

Profile    

Lityk

Post: 003  

26.03.2012 - 07:25

I was using following command line:
REM compile
C:\gcc\bin\gfortran.exe -c adapyukdriver.f fmmadapyuk.f prini.f second.f treeadap.f yukadap.f yukoperators.f yuktable.f 
REM link
C:\gcc\bin\gfortran.exe -g adapyukdriver.o fmmadapyuk.o prini.o second.o treeadap.o yukadap.o yukoperators.o yuktable.o -o fmm.exe

As you see I am on windows

Profile    

Zhangb

Post: 004  

06.06.2012 - 20:21

reason for including beta in the denominator

Hi Lityk,

Thanks for trying the code.

Please have a look at equation (2) of the paper "FMM-Yukawa: An Adaptive Fast Multipole Method for Screened Coulomb Interactions, published at Computer Physics Communication, 2009, vol. 180, 2331-2338".

That formula shows why we need to include lambda (called beta in the code) and also pi/2 in the calculation

Profile    

Please login/register to comment on this topic!

Page 1

This site has been visited 63693 times. Page: Forum.Program-ArticleDifference - Last Modified : Wed, Jun 06 2012

NSF-National Science Foundation DOE-US Dept of Energy

Maintained by Jun Jia, Bo Zhang, and Jingfang Huang
Copyright 2008 FastMultipole.org