Showing results for

- Intel Community
- Software Development SDKs and Libraries
- Intel® oneAPI Math Kernel Library & Intel® Math Kernel Library
- Sobol initialisation with direction numbers + primitive polynomials in fortran fails

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page

Highlighted
##

Dear experts,

could someone provide me with a working example for the initialisation of the sobol generator with both direction numbers and primitive polynomials in FORTRAN?

I am trying to use the dir. numbers and primitive polynomials from

http://web.maths.unsw.edu.au/~fkuo/sobol/new-joe-kuo-6.21201

by following the instructions

http://software.intel.com/sites/products/documentation/hpc/mkl/vslnotes/8_4_9_SOBOL.htm

but the random number sequence is the same no matter which direction numbers and prim. polynomials I use. Strangely enough, I don't get any error message about any problems during initialisation!

I am using a MKL version which came with fortran composer 2011.3.174.

Feedback is much appreciated.

Oleg

obrandt

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

11-29-2011
09:15 AM

10 Views

Sobol initialisation with direction numbers + primitive polynomials in fortran fails

could someone provide me with a working example for the initialisation of the sobol generator with both direction numbers and primitive polynomials in FORTRAN?

I am trying to use the dir. numbers and primitive polynomials from

http://web.maths.unsw.edu.au/~fkuo/sobol/new-joe-kuo-6.21201

by following the instructions

http://software.intel.com/sites/products/documentation/hpc/mkl/vslnotes/8_4_9_SOBOL.htm

but the random number sequence is the same no matter which direction numbers and prim. polynomials I use. Strangely enough, I don't get any error message about any problems during initialisation!

I am using a MKL version which came with fortran composer 2011.3.174.

Feedback is much appreciated.

Oleg

4 Replies

Highlighted
##

Gennady_F_Intel

Moderator

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

11-29-2011
09:36 PM

10 Views

Oleg,may be this example: " mrlroot\examples\vslf\source\osoboluserdirnums.f" will help you?

Highlighted
##

Dmitry_Petukhov__Int

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

11-30-2011
12:42 AM

10 Views

Oleg, let me give you a more detailed explanation.

1. SOBOL QRNG is implemented so that if params array is incorrectly filled the default implementation with dimension 1 will be used.

2. soboluserdirnums.f example shows how to initialize SOBOL QRNG with a complete direction numbers table. If you'd like you provide the generator with both initial direction numbers and primitive polynomials youll need to slightly change params array filling like this:

params(1) = dm1

params(2) = VSL_USER_QRNG_INITIAL_VALUES

params(3) = IOR(VSL_USER_INIT_DIRECTION_NUMBERS, &

& VSL_USER_PRIMITIVE_POLYMS)

params(3) = IOR(params(3), VSL_QRNG_OVERRIDE_1ST_DIM_INIT)

do i = 1, dm1

params(i+3) = poly(i)

end do

params(4 + dm1) = maxdeg

k = 5 + dm1

do i=1,dm1

do j=1,maxdeg

params(k) = vinit(i,j)

k = k + 1

end do

end do

nparams = 4 + dm1 + dm1*maxdeg

Highlighted
##

I did not know about this example!

sorry for not getting back to you earlier -- the autumn cold got me.

obrandt

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

11-30-2011
09:10 AM

10 Views

sorry for not getting back to you earlier -- the autumn cold got me.

Highlighted
##

Hi Dmitri,

thanks a lot for your more detailed explanation. At first glance it looks similar to my code, but I will have a close look once I'm back to work -- the autumn cold got me.

Best,

Oleg

obrandt

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

11-30-2011
09:14 AM

10 Views

thanks a lot for your more detailed explanation. At first glance it looks similar to my code, but I will have a close look once I'm back to work -- the autumn cold got me.

Best,

Oleg

For more complete information about compiler optimizations, see our Optimization Notice.