Re: Discussion summary regarding optimisations in 2D convolution during GSoC 2021.

From: Pranam Lashkari (plashkari628_at_[hidden])
Date: 2021-05-27 12:58:35

hello Prathamesh,

On Thu, May 27, 2021 at 1:51 PM Mateusz Loskot <mateusz_at_[hidden]> wrote:

>
> Generally, unless a new feature is aimed to be included in GIL as an
> extension,
> introduction of a new dependency on an external library should be avoided.
>
> Agreed here with Mateusz.

> If implementing functionality required by a feature is non-trivial,
> given time and scope of your project, then we need to consider
> splitting functionality of that feature into core part and extension part.
> Then, the extension part can rely on external library.
> Then, users will have choice, depending on their needs.
>
> Usually, my take on the complex task is to first implement the simplest
form of the feature, and then you keep on improving, looking at the time
frame this also seems to be better to do in this case. First to get simple
2d convolution/correlation up and running then we can take care of things
you mentioned in your first email one at a time. Starting simple gives a
better vision of the outcome we are targeting.

-- 
Thank you,
Pranam Lashkari, https://lpranam.github.io/