Extend 'Derivs' to cover 6th and 8th order finite difference

[lwJi]

Derivs:

1. implement 6th and 8th order accurate central finite difference.
2. implement dissipation and upwind derivatives corresponding to 6th and 8th order central finite difference.

TestDerivs:

1. implement unit test for all above.

all the coefficients used are calculated here using mathematica
https://github.com/lwJi/Docs-Subcycling/blob/main/wolfram/Derivs.ipynb

[rhaas80]

Fails in SYCL build due to calls via function pointer:

https://github.com/EinsteinToolkit/CarpetX/actions/runs/11491459037/job/31983862784?pr=307#step:6:5364

needs changes.

[rhaas80]

Fails due to compile for SYCL due to calls via function pointer:

https://github.com/EinsteinToolkit/CarpetX/actions/runs/11491459037/job/31983862784?pr=307#step:6:5364

probably also would fail (at runtime) for CUDA.

[lwJi]

Fails due to compile for SYCL due to calls via function pointer:

https://github.com/EinsteinToolkit/CarpetX/actions/runs/11491459037/job/31983862784?pr=307#step:6:5364

probably also would fail (at runtime) for CUDA.

I have remove the usage of function pointer, please take another look

[lwJi]

Hi @eschnett @rhaas80, I have addressed the comments in the above few commits. Thanks

[lwJi]

Hi @rhaas80, I think I have addressed all your comments in the recent few commits. please take another look. Thanks

[rhaas80]

Formatting change requested. Looks ok otherwise.

[lwJi]

Formatting change requested. Looks ok otherwise.

@rhaas80 Just pushed the formatting change. Thanks for catching it

[rhaas80]

Thank you. Looks good to me. You may want to reduce the (large) number of commits using a round of git rebase -i origin/main (makes it a bit easier in the future to see what changes happened).

[lwJi]

Thank you. Looks good to me. You may want to reduce the (large) number of commits using a round of git rebase -i origin/main (makes it a bit easier in the future to see what changes happened).

@rhaas80 , good point. I have squash them into 3 comments.