Add Python benchmarks for the new CUDA DSL/JIT

[ashvardanian]

Now that CCCL v3 can be used for efficient parallel reductions in Python it would be great to create an additional benchmark file - reduce_bench.py with Python-ic JIT-ed kernels for parallel reductions, showcasing the impact of different hyper-parameters on the result.

[AnshSinghSonkhia]

Hi @ashvardanian, I would like to work on this issue. Could you please assign it to me?

[ashvardanian]

Easy-peasy, consider it yours, @AnshSinghSonkhia! Keep me posted on the progress 😉