search

ExCALIBUR-NEPTUNE / NESO

MIT License
4 stars 4 forks source link

v0.1.0 #236

Closed will-saunders-ukaea closed 6 months ago

will-saunders-ukaea commented 6 months ago

Description

PR to stage v0.1.0 release in.

Type of change

Versioning.

Testing

Please do some - we have a variety of systems and it would be nice to find issues before making a release.

Changes

will-saunders-ukaea commented 6 months ago

TODO: update NS submodule for https://github.com/ExCALIBUR-NEPTUNE/NESO-Spack/pull/38

will-saunders-ukaea commented 6 months ago

NESO-Spack submodule has been updated - time to test.

will-saunders-ukaea commented 6 months ago

Spotted I mistakenly had numeric_limits::min instead of numeric_limits::lower - pushed update.

will-saunders-ukaea commented 6 months ago

I tried building the spack environment (gcc + hipsycl) and I am seeing:

     361    In file included from /home/js0259/git-ukaea/NESO/solvers/SimpleSOL/EquationSystems/SOLSystem.cpp:4:
  >> 362    /home/js0259/git-ukaea/NESO/solvers/SimpleSOL/EquationSystems/SOLSystem.hpp:6:10: fatal error: CompressibleFlowSolver/Misc/VariableConverter.h: No such file or d
            irectory
     363        6 | #include <CompressibleFlowSolver/Misc/VariableConverter.h>
     364          |          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

But when I do spack find -lv nektar the hash the spack env is using is +compflow_solver.

will-saunders-ukaea commented 6 months ago

The log reads:

==> Installing nektar-5.3.0-2022-09-03-ibp22bfpa7rovahfbmy7kftoyxko2fsr [43/45]
==> No binary for nektar-5.3.0-2022-09-03-ibp22bfpa7rovahfbmy7kftoyxko2fsr found: installing from source
==> Already patched nektar
==> nektar: Executing phase: 'cmake'
==> nektar: Executing phase: 'build'
==> nektar: Executing phase: 'install'
==> nektar: Successfully installed nektar-5.3.0-2022-09-03-ibp22bfpa7rovahfbmy7kftoyxko2fsr
  Stage: 0.00s.  Cmake: 7.34s.  Build: 7m 56.20s.  Install: 3.40s.  Post-install: 0.22s.  Total: 8m 7.50s
[+] /opt/spack-v0.21/linux-ubuntu22.04-cascadelake/gcc-12.3.0/nektar-5.3.0-2022-09-03-ibp22bfpa7rovahfbmy7kftoyxko2fsr

But the output directory suggests spack didn't patch nektar.

will-saunders-ukaea commented 6 months ago

Liberal application of git clean -fdx; git reset --hard HEAD to NESO and the nektar submodule seems to have solved this patch issue.

cmacmackin commented 6 months ago

Yes, this happens sometimes. Run rm .spack_patched and then build again. That should fix it.

will-saunders-ukaea commented 6 months ago

Tests passed on Archer2. Running into issues locally related to intel packaging rather than NESO.