www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.announce - Symmetry Autumn of Code 2026 Begins

reply Mike Parker <aldacron gmail.com> writes:
Two weeks ago, the SAOC 2026 judges sorted through 14 project 
proposals from 12 applicants.

We only provided one project idea this year and encouraged 
applicants to submit their own ideas. This resulted in the 
largest variety of project proposals we've ever seen. Only two 
applicants submitted the suggested project, and each of them 
submitted a backup project as recommended.

As usual, we have funding for only three slots. There were 
several good proposals didn't make it to the top of the combined 
rankings. These are the three that did, listed alphabetically by 
author name:

__ImportC: Zero-Wrapper Integration for Production Grade C 
Libraries__

Niyati Jain is working on this, with Nicholas Wilson as mentor. A 
number of ImportC bugs were squashed in SAOC 2025 by Emmanuel 
Nyarko. This project aims to build on that by targeting specific 
compilation failures in a set of widely-used C libraries and 
providing wrapper-free dub packages for those libraries. This 
involves fixing issues involving system headers, function-like 
macros, bitfields, and more.

__Improving DMD's CTFE Performance__

This one comes from Abul Khan, with Dennis Korpel as mentor. This 
project has two primary goals. The first is to make CTFE a 
replaceable component behind an interface. The second is to test 
it by taking a prototype modification Dennis implemented some 
time ago that improved memory usage and putting it behind the 
interface. This is not a full CTFE rewrite, but sets the stage 
for testing different engines later.

__MLIR Support for LDC__

Asadbek Sindarov porposed this, with Nicholas Wilson as mentor. 
LDC has unused stubs for Multi-Level Intermediate Representation 
(MLIR). This project will establish a functional, end-to-end MLIR 
backend for LDC. This will allow the utilization of advanced 
optimizations for D code and provide possibilities to leverage ML 
compiler framework optimizations.

Congratulations to Niyati, Abul, and Asadbek, and good luck with 
the event.

Thanks to everyone who submitted an application. I can tell you 
now that next year we will again skip a list of project ideas. If 
you think you might want to submit, I encourage you to start 
looking now for gaps you may be able to fill.

Thanks to Dennis and Nicholas for agreeing to take on the mentor 
role once again, and to Atila Neves, Robert Schadek, and Jonathan 
M. Davis for serving as our judges.

And a big thanks to Symmetry Investments for continuing to 
support this event every year. This is the 9th edition of SAOC 
already. Where does the time go?

Milestone 1 kicked off today. Keep an eye out for weekly reports 
in the General forum starting next week.
Sep 15
next sibling parent reply Serg Gini <kornburn yandex.ru> writes:
On Tuesday, 15 September 2026 at 14:55:03 UTC, Mike Parker wrote:

 __ImportC: Zero-Wrapper Integration for Production Grade C 
 Libraries__


Very nice! Proper dub integration would be much appreciated.
If there will be also capacity for SIMD/atomic/float8 support 
would be great!


 __Improving DMD's CTFE Performance__


I wonder if the approach was also influenced by Stefan's newCTFE?


 __MLIR Support for LDC__


This could make a great future proof advantage for D in 
heterogeneous computing!


 Congratulations to Niyati, Abul, and Asadbek, and good luck 
 with the event.


Congratulations! Good proposals and good luck with implementations


 Thanks to everyone who submitted an application.


Thanks to all involved!
Sep 15
parent Niyati Jain <niyati.jainn15 gmail.com> writes:
On Tuesday, 15 September 2026 at 14:55:03 UTC, Mike Parker wrote:

 Niyati Jain is working on this, with Nicholas Wilson as mentor. 
 A number of ImportC bugs were squashed in SAOC 2025 by Emmanuel 
 Nyarko. This project aims to build on that by targeting 
 specific compilation failures in a set of widely-used C 
 libraries and providing wrapper-free dub packages for those 
 libraries. This involves fixing issues involving system 
 headers, function-like macros, bitfields, and more.


Thank you so much, really appreciate the support. I already 
started working on it and will post updates on the forum as I go.

On Tuesday, 15 September 2026 at 15:23:55 UTC, Serg Gini wrote:

 Very nice! Proper dub integration would be much appreciated.
 If there will be also capacity for SIMD/atomic/float8 support 
 would be great!


Thanks! I'll try to fit SIMD/atomics in if there's time left 
after the core milestones, float8 is a bit out of scope rn for 
this term but I'll file it for later for sure and work on it.
Sep 17
prev sibling parent Hariprakash <hariprakash4389285 gmail.com> writes:
On Tuesday, 15 September 2026 at 14:55:03 UTC, Mike Parker wrote:

 Two weeks ago, the SAOC 2026 judges sorted through 14 project 
 proposals from 12 applicants.

 We only provided one project idea this year and encouraged 
 applicants to submit their own ideas. This resulted in the 
 largest variety of project proposals we've ever seen. Only two 
 applicants submitted the suggested project, and each of them 
 submitted a backup project as recommended.

 As usual, we have funding for only three slots. There were 
 several good proposals didn't make it to the top of the 
 combined rankings. These are the three that did, listed 
 alphabetically by author name:

 __ImportC: Zero-Wrapper Integration for Production Grade C 
 Libraries__

 Niyati Jain is working on this, with Nicholas Wilson as mentor. 
 A number of ImportC bugs were squashed in SAOC 2025 by Emmanuel 
 Nyarko. This project aims to build on that by targeting 
 specific compilation failures in a set of widely-used C 
 libraries and providing wrapper-free dub packages for those 
 libraries. This involves fixing issues involving system 
 headers, function-like macros, bitfields, and more.

 __Improving DMD's CTFE Performance__

 This one comes from Abul Khan, with Dennis Korpel as mentor. 
 This project has two primary goals. The first is to make CTFE a 
 replaceable component behind an interface. The second is to 
 test it by taking a prototype modification Dennis implemented 
 some time ago that improved memory usage and putting it behind 
 the interface. This is not a full CTFE rewrite, but sets the 
 stage for testing different engines later.

 __MLIR Support for LDC__

 Asadbek Sindarov porposed this, with Nicholas Wilson as mentor. 
 LDC has unused stubs for Multi-Level Intermediate 
 Representation (MLIR). This project will establish a 
 functional, end-to-end MLIR backend for LDC. This will allow 
 the utilization of advanced optimizations for D code and 
 provide possibilities to leverage ML compiler framework 
 optimizations.

 Congratulations to Niyati, Abul, and Asadbek, and good luck 
 with the event.

 Thanks to everyone who submitted an application. I can tell you 
 now that next year we will again skip a list of project ideas. 
 If you think you might want to submit, I encourage you to start 
 looking now for gaps you may be able to fill.

 Thanks to Dennis and Nicholas for agreeing to take on the 
 mentor role once again, and to Atila Neves, Robert Schadek, and 
 Jonathan M. Davis for serving as our judges.

 And a big thanks to Symmetry Investments for continuing to 
 support this event every year. This is the 9th edition of SAOC 
 already. Where does the time go?

 Milestone 1 kicked off today. Keep an eye out for weekly 
 reports in the General forum starting next week.


Congrats to Niyati, Abul, and Asadbek, all three of these sound 
like genuinely great projects. I submitted a proposal this year 
too (content-hash based build caching for dub) and while it 
didn't make the cut, I'm looking forward to following the 
progress here, especially the CTFE work, that's an area I've been 
circling around myself lately through some dub/DMD contributions. 
Good luck to all three of you this term.
Sep 18