improve: apply tachyon optimizations(2 & 5)

[guorong009]

Description

Analyze and apply the tachyon optimizations - 2(batch normalize) & 5(save heap allocations)

Related issues

• Analyze and select optimizations to port from C++ port of Halo2 by kroma-network/tachyon #293

Changes

• apply the batch_normalize in everywhere possible of halo2_backend crate

NOTE

1. batch_normalize
   In order to apply the batch_normalize optimization, the algorithm of batch_normalize & batch_inverse should be implemented for elliptic curves. It is already implemented in pse/halo2curves repo.

2. save heap allocations
   For this optimization, I just apply one - remove unnecessary "clone" op in "permutation_commit".
   The reason is that other optimizations needless in Rust project.
   Also, the Polynomial implementation has big diff between projects.

[ed255]

The changes LGTM!
I think the most significant one may be the memory optimization.
For the batch normalization, I believe the total cost saving is very small when done after an MSM because the MSM is already very costly (the MSM overshadows the improvement)