Re: FFT proposal for gsoc 2021

From: Eduardo Quintana (eduardo.quintana_at_[hidden])
Date: 2021-04-02 08:18:01

 ('
' encoding is not supported, stored as-is)
Hello,

Thanks to all the feedback that I have had in the previous days I've finished version
2.0 of the Gsoc proposal. That you can find here:
https://github.com/Lagrang3/gsoc-2021/releases/download/v2.0/gsoc.pdf

What's new?
- more realistic timeline
- less deriverables: no d-dimensional fft, no multi-threaded
- more focus on the API and back-ends
- included the abstract in the proposal
- definition of the convolution

I've also added some examples to the proof-of-concept git repository
https://github.com/Lagrang3/fftx

Let me stress that fftx's only purpose is experimentation and the competency test,
hence the API defined there is not final. I have not yet allocated time to make
that repository fully coherent with the GSOC proposal.

Best regards,
Eduardo

On Tue, Mar 30, 2021 at 07:00:11PM +0000, Christopher Kormanyos wrote:
> Hello Eduardo, Thank you for pursuing this project and preparing your powerful
> proposal. The proposal content its technical description, motivation and your
> personal details are very nicely and coherently presented.  These make your
> proposal already strong. As has been mentioned, a reduced scope of the project
> might be more realistic. And it seems like there is, in fact, strong interest
> in the interface.  Stefan mentions focus on an API that can be implemented as
> a wrapper around FFTW There is a lot of wisdom in this advice.  In fact, I
> encourage you also to look closely at our work with Boost.Multiprecision, one
> of my own projects co-authored John. There you will find that we provide both
> our own Boost-licensed multiple-precision types as well as wrapped versoins of
> GMP and MPFR. So in some sense, we have been living this advice for the past
> 10 years and it has been successful. I will comment again on your proposal in
> more depth. Feel free to refine it any way you like.  You might have to place
> it into the GSoC online format, as I am not sure if a manually-written LaTeX
> proposal can be used in a standalone form. But I'd have to check on that.
> Again, as mentioned and still the case, *I strongly encourage you to continue*
> with the application process. Kind regards, Chris