Hello, I would like to propose an idea for GSOC 2020. I would like to re-implement pre existing features in Boost-GIL for example inverting an image, conversion from black and white to colour and vice versa, etc, but with the usage of parallel programming in order to improve the efficiency and speed things up. This is nothing but an idea merely, as I have some idea with parallel programming beforehand. Could anyone give me some guidance as in how to proceed with this? And also do you all think this is a good enough proposal? I already have built a couple of libraries for image conversion using openmp, but I am not sure whether they are boost compatible or not. Thank You.
On Tue, 31 Dec 2019 at 16:33, Siddhartha Sen via Boost
I would like to propose an idea for GSOC 2020.
I believe I have already responded to very similar general request of yours: https://lists.boost.org/Archives/boost/2019/12/247771.php
I would like to re-implement pre existing features in Boost-GIL for example inverting an image, conversion from black and white to colour and vice versa, etc, but with the usage of parallel programming in order to improve the efficiency and speed things up.
In the thread I'm linking above, I also suggested that such implementation will have to be proposed as an extension. That is, for at least these reasons: 1. Parallel implementation must not replace existing GIL algorithms but should be proposed as complementary features. 2. GIL core library must not require OpenMP. 3. GIL extensions can require OpenMP.
Could anyone give me some guidance as in how to proceed with this?
I likely will repeat myself, but I'd suggest you to start with reading: 1. https://github.com/boostorg/gil/blob/develop/README.md 2. https://github.com/boostorg/gil/blob/develop/CONTRIBUTING.md and links from those two documents. Then, start designing, experimenting, prototyping, etc. i.e. the usual software craftsmanship applies. If you have any specific questions, if something doesn't build, if you get stuck, then post your questions here or boost-gil as recommended here: https://github.com/boostorg/gil/blob/develop/README.md#community
And also do you all think this is a good enough proposal?
This is not a proposal, but just an idea tossed into the air. Please, read the GSoC guides to learn how to write a proposal https://google.github.io/gsocguides/student/ I'd also point at this article which every student interested in GSoC should read, IMHO: https://github.com/boostorg/boost/wiki/Google-Summer-of-Code%3A-Advice-From-... You are also welcome to implement your idea and submit to GIL as a new feature via Pull Request (PR) before even GSoC starts. Every Boost library accepts PRs proposing new features.
I already have built a couple of libraries for image conversion using openmp, but I am not sure whether they are boost compatible or not.
I don't understand what kind of compatibility you mean. Perhaps you should read these https://www.boost.org/users/faq.html https://www.boost.org/development/requirements.html and see if anything there addresses your concern. Best regards, -- Mateusz Loskot, http://mateusz.loskot.net
participants (2)
-
Mateusz Loskot -
Siddhartha Sen