v0.20.16

MPI v0.20.16

Diff since v0.20.15

Merged pull requests:

• Fix GPU tests (#764) (@luraess)
• Cleanup and update doc (#765) (@luraess)
• Add non-blocking wait (#766) (@luraess)
• Bump AMDGPU compat (#767) (@luraess)
• Bump actions/checkout from 3 to 4 (#768) (@dependabot[bot])
• improve docs on modulefiles (#769) (@simonbyrne)
• Init cooperative wait test for multiple threads (#770) (@lcw)
• Fix resize! for (Unsafe)MultiRequest (#772) (@jipolanco)
• Docs: update to Documenter v1 (#773) (@jipolanco)
• Bump CUDA compat (#774) (@simonbyrne)
• Bump patch (#775) (@simonbyrne)