Release 3.0

[arporter]

As discussed, we need to make a 3.0 release. Mainly to facilitate the new users of the 'code transformation' functionality and to emphasise that the API has changed since the 2.5.0 release.

• (Closes #320) nonstandard stop fparser#451
• 448 make public example fparser#450
• Loop fusing and openmp transformations generating incorrect code #2701
• FFSL support is required (FFSL transport in LFRic depends on a kernel that loops over halo cells #2661)
• Document the OpenMP offload functionality (under https://psyclone.readthedocs.io/en/latest/transformations.html#openmp) and add/extend an LFRic example to show it.
• Update LFRic transformation signature? #2739
• [LFRic] [PSyAD] Function space pointers declared and assigned in random orders in generated adjoint test #2742 (PSyAD test-harness reproducibility)
• [LFRic] [PSyAD] PSyAD no longer computes adjoint of vector assigments properly in PSyclone 3.0-dev #2755
• (towards #2671) Parallel array privatisation #2697 Would be nice to have as it also fixes NEMO integration tests.
• Update version in Spack (and update Wiki to remind us to do this)

We want this before SC24.

[hiker]

• 2690 value verification #2694
• Loop Fusion Transform can create incorrect code #257

I might also consider #2483, that might be useful for work with the UM, but I need to check where this PR is exactly at

[DrTVockerodtMO]

• (Closes 2311) Add functionality to derive module and subroutine names of adjoint test from filename for LFRic API #2749

This task is almost complete so I have added it to the list.

[LonelyCat124]

@arporter Should we rename the "nemo" tutorials to Code transformation before this release as well?

[arporter]

@arporter Should we rename the "nemo" tutorials to Code transformation before this release as well?

Ideally, yes.

[LonelyCat124]

@arporter Should we rename the "nemo" tutorials to Code transformation before this release as well?

Ideally, yes.

I'll see if I can do this then

[hiker]

@arporter Should we rename the "nemo" tutorials to Code transformation before this release as well?

Ideally, yes.

I'll see if I can do this then

#2759 (which I assume should go into 3.0 as well) addresses a similar issue with library files. And there are the examples as well.

We should agree on a name - maybe just transformation? Or no_psykal?

[sergisiso]

For the folder names: generic or std_fortran are other options.