Make QMC_SIMD_ALIGNMENT configured via CMake

[ye-luo]

Proposed changes

Renamed QMC_CLINE to QMC_SIMD_ALIGNMENT.
Set QMC_SIMD_ALIGNMENT (SIMD vector length in bytes, not necessarily cache line size) explicitly via CMake rather than depending on compiler defined macros.
The default value is determined by check __AVX512F__ macro. If found, set 64 otherwise 32.
Basically the compiler macro logic is moved to CMake and the value is configurable but checked in CMake.

Background:
The old code was fine when targeting only one architecture (CPU).

  #if defined(__AVX512F__)
    #define QMC_SIMD_ALIGNMENT 64
  #else
    #define QMC_SIMD_ALIGNMENT 32
  #endif

When OpenMP offload is used, the source code is compiled for two targets CPU and GPU.

int foo(Vector<float, aligned_allocator<float, QMC_SIMD_ALIGNMENT>>)
{
  #pragma omp target
  { offload kernel }
}

The compiler transformed offload kernel carries C++ mangled function name foo with QMC_CLINE being part of it. For this reason, the caller (CPU pass) and the callee(GPU pass) doesn't match and kernels failed being loaded on GPU. https://bugs.llvm.org/show_bug.cgi?id=45401

What type(s) of changes does this code introduce?

• Bugfix

Does this introduce a breaking change?

• No

What systems has this change been tested on?

Bora, Summit, Tulip, Theta

Checklist

• Yes. This PR is up to date with current the current state of 'develop'

[prckent]

1. Please change QMC_CLINE to a self-describing and longer name. e.g. QMC_SIMD_ALIGNMENT. What CLINE means is not obvious.

2. Rename CheckSIMDAlignement.cmake ->CheckSIMDAlignment.cmake

[prckent]

Please add an example of the potential problem, as detailed in your PR text.

[prckent]

Thanks Ye!

[ye-luo]

@prckent This should make the nightly test error on sulfur go away without need of passing "-march=skylake" as a workaround.

[prckent]

Nightly scripts have been updated.