Dear Sirs or Madams:
I am sorry for my question - I realize it is probably fairly simple,
however I have been wading through the BLAS/LAPACK/SCALAPACK/PLAPACK
landscape for a bit and am still quite lost. Any direction will be much
appreciated.
I am interesting in using SCALAPACK to parallelize drawing from a large
multivariate normal distribution (60,000+ variables). I understand
that the Cholesky factorization is the most time-intensive step, and,
further, I understand that SCALAPACK has parallelized this.
Thanks to your wonderful installer, I have managed to successfully set
up SCALAPACK on our cluster. Now, I would like to pursue
the next step -- implementing the multivariate normal sampler using
SCALAPACK, preferably in C as I am quite novice at Fortran.
Any hints in this direction would be much appreciated. I am trying to
look through your Website. It is very informative and it
seems like your package can do so many nice things, but I am a little
lost as to how to accomplish my rather narrow and focused goal.
Thank you so much.
Sincerely yours
Misha Koshelev
--
Misha Koshelev
MD/PhD Student
Human Neuroimaging Laboratory
One Baylor Plaza
S104
Baylor College of Medicine
Houston, TX 77030
|