v2023.06.0 Release

[rhornung67]

Items to address before the release:

• Update GitLab CI so that it runs on ruby. This Update RAJA submodule and CI #347 + follow-on PRs as needed.
• Update RAJA, includes multiple PRs
  • Update RAJA to v2023.06.1 release
  • Make GitLab CI jobs consistent with RAJA, including:
    • Switch GitLab CI to service user account
    • Add +omptask build variant where appropriate
• Merge Copy8 kernel: Add COPY8 kernel #337
• Merge Zonal Accumulation kernel: Add ZONAL_ACCUMULATION_3D kernel #336
• Merge FEM mass matrix assembly kernel: FEM Mass matrix element assembly kernel #330
• Merge EDGE3D kernel: EDGE3D App #300
• Merge macro spelling fix: Fix spelling of macro name #345
• Merge tunings command line option: PR from fork/333 #335
• Merge Caliper Integration: PR from fork/331 #352
• Merge MPI halo exchange: Add MPI HALOEXCHANGE kernels #332
  -- note: we need to add MPI builds to CI checks
• Merge kernel that captures array and length: Add kernel that captures an array and length #258
• Change loop_exec policies to seq_exec: Change all loop_exec policies to seq_exec after RAJA release is merged in #341
• Convert GPU exec to use new policies: New RAJA GPU Indexing #342
• mat-mat shared checksum issue: MAT_MAT_SHARED checksums are bad #325
• Improve warm-up kernels #312
• Dispatch policies for halo exchange: HALOEXCHANGE_FUSED WorkGroup Dispatch Policies  #260
• Same GPU stream for all kernels: Use same GPU stream for all kernels #296

Others???

• script to make sweep graphs: Add script to make graphs from problem size sweeps #245
• HIP MFMA checks: hip mfma tests #246 (need to fill out other back-ends?)
• Unsafe atomic tunings for HIP: [Draft] Refactor GPU reductions and add unsafe atomic tunings #247

[rhornung67]

Release is done