[FEA] Reduce libcuopt binary size

[rgsl888prabhu]

Is your feature request related to a problem? Please describe.
As of 25.08, the libcuopt wheels have the following sizes:

distribution	arch	size (compressed)	size (uncompressed)
libcuopt-cu12	x86_64	0.425G	0.579G
libcuopt-cu12	aarch64	0.425G	0.593G

These details have been fetched from CI run for nightly packages.

Describe the solution you'd like

The package size should be reduced as much as possible,

PyPI requirements for individual files:

over 100 MiB = requires an exception (see https://github.com/pypi/support/issues)
over 1GiB = absolutely not allowed

Need to go over the wheel package and remove any static linking and move them to dynamic and also try to optimize binary.

[anandhkb]

PR to address this in progress via #507