Updating to SCORPIO v1.6.7

[jayeshkrishna]

SCORPIO v1.6.7 includes the following fixes,

• Fix for E3SM restart issues with ADIOS
• CMake fixes (reduces warnings, workaround for CMake 3.22 bug)
• Support for ADIOSC in legacy Fortran interface

[BFB]

[bartgol]

Need an snl-testing member approval to run EAMxx tests

[rljacob]

@jayeshkrishna the CI testing shows some fails for eamxx.

[jayeshkrishna]

@bartgol : Do you think the eamxx fails above are related to SCORPIO? I see a lot of EKAT assertion fails like,

/home/runner/_work/E3SM/E3SM/externals/ekat/src/ekat/kokkos/ekat_subview_utils.hpp:33: Kokkos::View<Kokkos::View<ST *, Kokkos::LayoutRight, Props...>::traits::scalar_array_type, Kokkos::View<ST *, Kokkos::LayoutRight, Props...>::traits::array_layout, Kokkos::View<ST *, Kokkos::LayoutRight, Props...>::traits::device_type, Kokkos::MemoryTraits<<expression>>> ekat::subview(const Kokkos::View<ST **, Kokkos::LayoutRight, Props...> &, int) [with ST = const float; Props = Kokkos::Device<Kokkos::Cuda, Kokkos::CudaSpace>; Props = Kokkos::MemoryTraits<0U>]: block: [124,0,0], thread: [0,0,0] Assertion `i0>=0 && i0 < v.extent_int(0)` failed.

[mahf708]

I don't think the errors in the CI are related to SCORPIO; they seem to be related to other known errors (mam wetscav, microphys, etc.)

[bartgol]

Yeah, I think the errors came from an old bug in mam, which got recently resolved. I will trigger a new set of testing to be sure.

[jayeshkrishna]

@bartgol : The test cosp_standalone_baseline_cmp is failing in the EAMXX test checks below, the fail is a DIFF in comparison with baselines. Is this related to this PR (or is a separate known issue)?

[bartgol]

@jayeshkrishna there was a pr that fixed cosp that went in last week. I think I forgot to bless master until Monday. I'm rerunning the failed tests, in case the CI for this PR ran before the blessing on Monday. Will know soon enough.

Edit: looks like they all pass now. Good to go on the eamxx side.