epic: llama.cpp should support `sycl` for Intel-based CPUs

[dan-menlo]

Goal

• Should support sycl for Intel-based CPUs

Tasklist

• What should be the hardware heuristic to opt for LLVM?
• Should Installer support this?
• This may require us to have cortex engines be able to define a "default" llama.cpp version (if two or more exist)

[dan-menlo]

Closing to handle as part of #1453