[gsoc21] - Provide quad-double precision for Boost.Multiprecision
Hello, My name is Giannis Gonidelis. I am an Electrical & Computer Engineering student from Greece and a regular open-source contributor on HPX developed by STE||AR Group. The past year I have been working on trying to adapt the HPX Parallel Algorithms to C++20 by providing the interface dictated by Ranges-TS (by Niebler). HPX, being a C++ Standards library for Concurrency and Parallelization, has its roots on Boost and widely adopts template meta-programming (our project leader is also the founder of Boost.Spirit). I recognize that I might kick in a little late for a GSoC application, but before sharing my proposal, I would like to know if you plan on mentoring the project about providing the "Quad-Double Backend Type". If so, I would be glad to share my initial thoughts on the project with the mentor, along with my Programming competency test script, in order to submit a cohesive and thorough proposal. Please inform me on the project availability. Regards, Giannis Gonidelis GitHub: gonidelishttps://github.com/gonidelis
Dear Giannis,
Thank you for your query and yourinterest in "provide quad-double".
The project is open. You are right on time!
I would be the mentor (Christopher Kormanyos).At the moment, I have two potential projects.I will, however, have time to mentor just 1 ofthem, either this one or the FFT proj.
I strongly encourage you, if your scheduleallows, to submit your proposal and try.
Kind regards, Chris
On Thursday, April 8, 2021, 8:16:13 PM GMT+2, Γιάννης Γονιδέλης via Boost
Dear Giannis,
Thank you for your query and your interest in "provide quad-double".
The project is open. You are right on time!
I would be the mentor (Christopher Kormanyos). At the moment, I have two potential projects. I will, however, have time to mentor just 1 of them, either this one or the FFT proj.
I strongly encourage you, if your schedule allows, to submit your proposal and try.
Thank you for the information! I have submitted my Draft Proposal to the GSoC portal where I fully explain my implementation schedule and my C++ background. Please, feel free to comment and make suggestions. It would help a lot in order to form my final proposal.
Kind regards, Chris
Regards, Giannis Gonidelis
Dear Giannis,
I find your proposal very good, showingclear understanding of the topic, a realistictimeline and presents you overall asa very competent and confident candidate.
Good job.
Optionally...
You might consider in the first paragraphof Chapter 2 in your proposal, you mentionthe characteristics of QD. In the secondparagraph, you also mention benching it.I think in the first paragraph, you couldpotentially mention that QD is expectedto offer higher performance in its digitrange than existing Boost.Multiprecisionbackends in this digit range.
Kind regards, Christopher Kormanyos
On Saturday, April 10, 2021, 11:04:23 PM GMT+2, Γιάννης Γονιδέλης
Thank you for your query and your> interest in "provide quad-double".> The project is open. You are right on time!> I would be the mentor (Christopher Kormanyos).> At the moment, I have two potential projects.> I will, however, have time to mentor just 1 of> them, either this one or the FFT proj.>> I strongly encourage you, if your schedule> allows, to submit your proposal and try.> Thank you for the information! I have submitted my Draft Proposal to the GSoC portal where I fully explain my implementation schedule and my C++ background. Please, feel free to comment and make suggestions. It would help a lot in order to form my final proposal. Kind regards, Chris Regards,Giannis Gonidelis
Dear Giannis,
I find your proposal very good, showing clear understanding of the topic, a realistic timeline and presents you overall as a very competent and confident candidate.
Thank you very much for taking a look. I appreciate it!
Good job.
Optionally... You might consider in the first paragraph of Chapter 2 in your proposal, you mention the characteristics of QD. In the second paragraph, you also mention benching it. I think in the first paragraph, you could potentially mention that QD is expected to offer higher performance in its digit range than existing Boost.Multiprecision backends in this digit range.
I have taken your word into consideration and I pushed several fixes on that (GSoC portal). I would be glad if you could share some final thought before final submition. I would also like to inform you that I will enhance my final submission by providing a Multiprecision use case example on the "Bisextion Method". It's not there linked to my application because I want to fix some last details, but you could for sure review the remaining document.
Kind regards, Christopher Kormanyos
I would be glad to work with you during the summer. Regards, Giannis Gonidelis
share some final thought before final submition... Hi Giannis, Please mark your proposal as *final* today,since today is the deadline to do so. Your proposal is a very good proposal,and even better and more complete nowwith your changes. I repeat: Please mark your proposalas final today.
Kind regards, Chris
On Tuesday, April 13, 2021, 12:15:20 AM GMT+2, Γιάννης Γονιδέλης
I find your proposal very good, showing> clear understanding of the topic, a realistic> timeline and presents you overall as> a very competent and confident candidate. Thank you very much for taking a look. I appreciate it! Good job.> Optionally... You might consider in the first paragraph> of Chapter 2 in your proposal, you mention> the characteristics of QD. In the second> paragraph, you also mention benching it.> I think in the first paragraph, you could> potentially mention that QD is expected> to offer higher performance in its digit> range than existing Boost.Multiprecision> backends in this digit range. I have taken your word into consideration and I pushed several fixes on that (GSoC portal).I would be glad if you could share some final thought before final submition.I would also like to inform you that I will enhance my final submission by providinga Multiprecision use case example on the "Bisextion Method". It's not there linkedto my application because I want to fix some last details, but you could for sure review the remaining document.
Kind regards, Christopher Kormanyos
I would be glad to work with you during the summer. Regards,Giannis Gonidelis
participants (2)
-
Christopher Kormanyos -
Γιάννης Γονιδέλης