Re: [petsc-users] MPI-FFTW example crashes

[Smith, Barry F. via petsc-users]

> On Jun 2, 2019, at 11:52 PM, Sajid Ali <sajidsyed2...@u.northwestern.edu> 
> wrote:
> 
> @Barry : Perahps config/BuildSystem/config/packages/fftw.py should use the 
> --host option when configure for PETSc is run with-batch=1. 
> 
> If anyone here knows what --host must be set to for KNL, I'd appreciate it. 

  We've tried very hard over the years to avoid learning about the --host 
business. I found a couple of useful pages

https://stackoverflow.com/questions/21990021/how-to-determine-host-value-for-configure-when-using-cross-compiler

https://www.gnu.org/software/autoconf/manual/autoconf-2.65/html_node/Specifying-Target-Triplets.html

I don't know if the host stuff will work on theta since it is such a unique 
machine.

 Good luck

> 
> PS :  I know that Intel-MKL-FFT provides FFTW api. If I'd want to try with 
> this, is there a way to tell PETSc to pick fftw functions from MKL ? 

 We've never done it but take a look at mkl_sparse.py perhaps it could be 
converted to mkl_fftw.py ?

> 
> 
> Thank You, 
> Sajid Ali
> Applied Physics
> Northwestern University