fix: optimize compute_inner_product with parallel computation

[crStiv]

Description:

This PR optimizes the compute_inner_product function by implementing parallel computation for large vectors while maintaining efficient sequential computation for small ones.

Changes:

• Implemented parallel computation for vectors larger than 32 elements
• Kept sequential computation for small vectors to avoid parallelization overhead
• Added comprehensive documentation explaining the optimization strategy
• Added tests covering both sequential and parallel computation paths
• Removed TODO comment as parallelization is now implemented

The optimization uses the existing parallelize infrastructure to split the computation into chunks and process them in parallel, which should improve performance for large vectors while maintaining efficiency for smaller ones.

Tests:

• Added test cases for both small (16 elements) and large (64 elements) vectors
• Tests use random data to ensure correctness
• Tests verify results against sequential computation

This change maintains backward compatibility and follows the project's coding style.

[jonathanpwang]

Use rayon sum instead for better performance.

[crStiv]

@jonathanpwang this one is alright?

[jonathanpwang]

@jonathanpwang this one is alright?

your last commit seems to be empty

[crStiv]

@jonathanpwang I dunno what's wrong with the previous one, but now it seems to be all alright

[jonathanpwang]

LGTM

[jonathanpwang]

Lints broken

[crStiv]

Lints broken

@jonathanpwang
sorry for taking much of your time, but checks are green now and probably now everything should be fine

[jonathanpwang]

Need to include rayon = 1.8 in Cargo.toml
I will just merge and fix in followup