Package: libmathlib1
Version: 2004.11.04-1
There is some (probably compilation) bug in the libmathlib1 package,
that arises when using the subrutine DADMUL for multidimensional
integration.
Please, consider the following program
program testdadmul
implicit none
real*8 A(3),B(3),WK(150000),pi,valor,relerr,eps,c(3),t
integer NFNEV,IFAIL,IWK,minpt,maxpt
external f3
c double precision f3
pi=dacos(-1.d0)
A(1)=0.d0
A(2)=0.d0
A(3)=0.d0
B(1)=pi
B(2)=pi
B(3)=pi
c(1)=1.d0
c(2)=1.d0
c(3)=1.d0
IWK=150000
eps=1.d-3
minpt=64
maxpt=1000000
c t=f3(3,c)
write(*,*) 'calling dadmul'
CALL DADMUL(f3,3,A,B,minpt,maxpt,eps,WK,IWK,valor,relerr,NFNEV,
. IFAIL)
write(*,*) 'I=',valor,'+-',relerr,'%',NFNEV,' evals',' error
. ',IFAIL
end
real*8 function f3(Ndim,x)
implicit none
double precision x(3),a,b,c
integer Ndim
a=x(1)
b=x(2)
c=x(3)
f3=(sin(a)*sin(b)*sin(c))**2
return
end
I compile it with the command
g77 -static -o test.x test.f -lmathlib -lpacklib -lkernlib
and running it, I get the following output
./test.x
calling dadmul
I= 3.87581798+- 0.000980121412% 38115 evals error 0
The 'I' above is the value of the integral. The exact value is pi^3/8
However, if I change the lower integration limits (0.d0 originally) to
A(1)=1.d-8
A(2)=1.d-8
then I get this result
g77 -static -o test.x test.f -lmathlib -lpacklib -lkernlib
./test.x
calling dadmul
I= 42336.3332+- 2.45015127% 999999 evals error 1
The I value here is completely wrong.
This is my system:
[EMAIL PROTECTED]:~$ cat /proc/version
Linux version 2.6.8-1-686 ([EMAIL PROTECTED]) (gcc version 3.3.5 (Debian
1:3.3.5-2)) #1 Thu Nov 25 04:34:30 UTC 2004
I have compiled with the versions 3.3.4 and 2.95 of g77, and
also tested it in a computer with Mandrake linux.
Then, I downloaded from the CERN the source file containing the subroutine
radmul/dadmul, and compiled it in the following way
g77 -g -static -c -o inttegint.o inttegint.f
g77 -g -static -c -o radmul.o radmul.f
g77 -g -static -o a.x driver.f inttegint.o radmul.o -lmathlib -lpacklib
-lkernlib
Now the program runs OK.
./test.x
calling dadmul
I= 3.87578459+- 2.00515818E-16% 495 evals error 0
Thank you
Eduardo A. Menendez Proupin
Department of Physics
Faculty of Science
University of Chile
Las Palmeras 3425
�u�oa, Santiago
Chile
Phone: 56+2+678 74 11
http://fisica.ciencias.uchile.cl/~emenendez/
