Improve the support of running GS calculations parallel in states

Description

Running GS calculations parallel in states has been marked experimental. This MR improves the support of running parallel in states for the CPU and GPU versions. Right now, this is slower than running parallel in domains, but still it works in case it is needed for weak scaling. These improvements are for the implementation not using Scalapack.

News snippet

Improve the support of running GS calculations parallel in states

Checklist

• I have checked that my code follows the Octopus coding standards
• I have added tests for all the new features added in this request.