Add range minimum query kernels

[upsj]

This adds a data structure to answer queries of the kind $\text{argmin}_{i \in [a,b]} \text{array}[i]$ in constant time. This will be used to speed up the computation of LCAs in the symbolic Cholesky algorithm, turning the counting step from $O(nnz(L))$ algorithm to a $O(nnz(A))$ algorithm.

The corresponding Wikipedia articles describe the algorithms pretty well https://en.wikipedia.org/wiki/Range_minimum_query, the constant time query is described in more detail in J. Fischer and V. Heun, “Space-Efficient Preprocessing Schemes for Range Minimum Queries on Static Arrays,” SIAM J. Comput., vol. 40, no. 2, pp. 465–492, Jan. 2011, doi: 10.1137/090779759.

TODO:

• add test for device queries

[yhmtsai]

my first impression on turning $O(nnz(L))$ to $O(nnz(A))$ is more operation.
do I miss something?

[upsj]

L includes fill-in, A does not

[MarcelKoch]

Overall, a lot more documentation is necessary. Especially what the new classes in bit_packed_storage and range_minumum_querry are there for, and how they work together.
At least I can't properly review it right now, since there is too little information on what functions/classes are supposed to do.

[MarcelKoch]

I would like to have the discussion on the description of lookup kernels resolved. The rest are smaller nits.

[MarcelKoch]

Is it possible to describe in a few words the meaning of C(p, q) in this context?

[upsj]

I don't think there is any particularly useful description I could add here, beyond the definition above.

[sonarqubecloud]

Quality Gate passed

Issues
81 New issues
0 Accepted issues

Measures
0 Security Hotspots
98.6% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud