Use mpi_core_exp_mod in bignum

[yanesca]

Description

Use mpi_core_exp_mod() in bignum.

The two algorithms are not equivalent. The original bignum exponentiation was a sliding window algorithm. The one in mpi_core_exp_mod() uses a fixed window approach. This change is intentional. We don't want to maintain two algorithms and decided to keep the fixed window algorithm.

The change of algorithms impacts performance. To keep the performance comparable to previous versions, the window size had to be increased.

The choice of the parameter was based on measurements on my development machine (Intel i5). This is 2nd-best and 2nd-worst of 5 runs, for three key sizes, in operations per second (so, higher is better). (As in https://github.com/Mbed-TLS/mbedtls/issues/6790.)

w	6	5	4	3	2	1	dev
2048	392-398	421-426	426-433	426-432	398-401	286-317	441-456
3072	148-149	152-158	145-152	150-153	133-136	108-108	143-156
4096	71-72	72-73	69-73	61-70	63-64	49-49	72-72

Fixes #6600 .

PR checklist

Please tick as appropriate and edit the reasons (e.g.: "backport: not needed because this is a new feature")

• [ ] changelog provided
• [ ] backport not required
• [ ] tests already present