[LLVM][refactoring] Preparing code for builders

[georgemitenkov]

Since #911 was a huge change and difficult to review, it is better to split into multiple PRs. This one reorganises code, fixes some typos ad renames functions. Some code has now explicit if (GPU) else if (SIMD) else blocks which is what we need to have multiple builders.

[bbpbuildbot]

Logfiles from GitLab pipeline #92357 (:no_entry:) have been uploaded here!

Status and direct links:

• ✅ spack_setup
• ✅ build:intel
• ✅ build_cuda:gcc
• ✅ test:intel
• ⛔ test_benchmark:gcc

[iomaganaris]

FYI the Azure CI are failing due to incompatibilities between the llvm branch of NMODL and the latest version of NEURON being used. I would only focus on the GitLab CI that runs the CPU and GPU integration tests of the JIT workflow

[bbpbuildbot]

Logfiles from GitLab pipeline #92578 (:white_check_mark:) have been uploaded here!

Status and direct links:

• ✅ spack_setup
• ✅ build:intel
• ✅ build_cuda:gcc
• ✅ test:intel
• ✅ test_benchmark:gcc