Hi,
This is the status report for the GSoC project that I'm still working
on. Sorry for the batch of time that I didn't send a work report.
This time I set out to:
- Work in the bind() clause implementation.
- Keep an eye in the review of the Forge PR for the C and C++
OpenACC directives support
Over this time:
- I re-configured my setup and refactor the build utility that I
have to support a configuration file to make my life easier.
- I got some input for the patch that I sent in Forge (C and C++
OpenACC directives support) and fix them.
- I sent the PR to Forge adding the support for init,
set and shutdown to the C and C++ frontends
https://forge.sourceware.org/gcc/gcc/pulls/225
- I got some comments from my mentors in that PR but I forgot to follow
those instructions (separate the patch into 3 patches).
- I implemented a prototype for the bind() clause, and get some
feedback from my mentors in the implementation.
Current status
- I just sent the patch series to the mailing list following the
instructions from my mentors:
https://gcc.gnu.org/pipermail/gcc-patches/2026-October/733396.html
https://gcc.gnu.org/pipermail/gcc-patches/2026-October/733395.html
https://gcc.gnu.org/pipermail/gcc-patches/2026-October/733394.html
I canceled the PR from Forge because I consider easier to send
those patches to the mailing list directly, since I had to split
the commit into three.
The next patch I'll test Forge workflow again!
- The bind() clause implementation got some errors from the
testcases that my mentors suggests, this two weeks I was
experimenting a lot with GIMPLE, specifically with the attributes
and how the IR translates to GIMPLE in some moment to solve the
issue with the clause implementation.
The priority is to catch up the bind() cluase patch and continue with
the next task (device_type implementation), also reviewing the OpenACC
spec looking for unimplemented features in order to add support in GCC
to those features.
Sorry again for the delay!
Best,
Sebastian
