<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Wed, Aug 20, 2014 at 4:05 PM, Barry Smith <span dir="ltr"><<a href="mailto:bsmith@mcs.anl.gov" target="_blank">bsmith@mcs.anl.gov</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class=""><br>
On Aug 20, 2014, at 4:00 PM, Matthew Knepley <<a href="mailto:knepley@gmail.com">knepley@gmail.com</a>> wrote:<br>
<br>
> On Wed, Aug 20, 2014 at 3:52 PM, Barry Smith <<a href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a>> wrote:<br>
><br>
> On Aug 20, 2014, at 3:34 PM, Matthew Knepley <<a href="mailto:knepley@gmail.com">knepley@gmail.com</a>> wrote:<br>
><br>
> > The blurb should also answer the question, Why do we need another FFT library?<br>
><br>
> Because it actually does multidimensional FFTs in parallel?<br>
><br>
> For something this old and established, it might be important to say that it does blah<br>
> blah blah that Spiral, FFTW, etc.<br>
><br>
> In theory, with this one could write very efficient parallel 3d Poisson solvers in PETSc for boxes, which is an important special case that PETSc does not currently support.<br>
><br>
> Wouldn't you just use MG?<br>
<br>
</div> FFT when done properly is much faster! Yes it is a specialized case but an important one.</blockquote><div><br></div><div>Are we calculating the following way:</div><div><br></div><div> 1) MF MG for the Laplacian is about 50 F/dof</div>
<div><br></div><div> 2) An FFT is about 6 F/dof, and we use 1 transform, 1 divide, and 1 inverse transform, so we get 13 F/dof</div><div><br></div><div>so you get 3x or so?</div><div><br></div><div> Matt</div><div> </div>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">
> I have other uses for FFT, including DFT codes where it makes a lot of sense.<br>
><br>
> Matt<br>
><br>
><br>
> Barry<br>
><br>
> ><br>
> > Matt<br>
> ><br>
> ><br>
> > On Wed, Aug 20, 2014 at 1:53 PM, Barry Smith <<a href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a>> wrote:<br>
> ><br>
> ><br>
> > From: Dmitry Pekurovsky <a href="mailto:dmitry@sdsc.edu">dmitry@sdsc.edu</a><br>
> > Date: August 12, 2014<br>
> > Subject: Library for spectral transforms in 3D for parallel machines<br>
> ><br>
> > P3DFFT is an open-source numerical library providing highly<br>
> > scalable implementation of 3D spectral transforms namely Fast<br>
> > Fourier Transform (FFT), with an option to combine it with<br>
> > cosine/sine/Chebyshev/empty transform in the third dimension. (The<br>
> > empty transform allows the user to substitute their own custom<br>
> > transform in the third dimension. This can be useful in<br>
> > applications such as inhomogeneous wall bounded turbulence.) P3DFFT<br>
> > implements 2D domain decomposition which allows it to overcome a<br>
> > scalability restriction inherent in 1D decomposition. This approach<br>
> > has shown good scalability up to 131,072 cores.<br>
> ><br>
> > A new version of P3DFFT 2.7.1 is now available. The project Home<br>
> > Page is <a href="http://code/google.com/p/p3dfft" target="_blank">http://code/google.com/p/p3dfft</a> where instructions for<br>
> > obtaining the source code are provided. Installation instructions<br>
> > and a User Guide are also available.<br>
> ><br>
> > P3DFFT features include real-to-complex and complex-to-real<br>
> > transforms, in-place transforms, pruned transforms (with less than<br>
> > full input or output), and multi-variable transforms. The package<br>
> > includes example programs in Fortran and C. This is a project in<br>
> > active development, with a user mailing list, a wiki page and a<br>
> > version control system. P3DFFT is considered community software and<br>
> > is being installed in public space at many supercomputer centers.<br>
> > Contributions and feedback from users are welcome.<br>
> ><br>
> ><br>
> ><br>
> > --<br>
> > What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
> > -- Norbert Wiener<br>
><br>
><br>
><br>
><br>
> --<br>
> What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
> -- Norbert Wiener<br>
<br>
</div></div></blockquote></div><br><br clear="all"><div><br></div>-- <br>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
-- Norbert Wiener
</div></div>