Use `shuffle_list` instead of `compute_shuffled_index` when possible.

[mpaulucci]

Preliminary read: https://eth2book.info/capella/part2/building_blocks/shuffling/#a-full-shuffle
Full shuffle implementation: #467

Using compute_shuffled_index is quite slow, so instead of computing it for every index that we care about. Instead we should shuffle the whole list once and then query each index. Maybe we'll need some type of cache for this.

Some places where this function is used

lambda_ethereum_consensus/lib/lambda_ethereum_consensus/state_transition/misc.ex

Line 207 in e01f275

{:ok, index} = compute_shuffled_index(i, index_count, seed)

lambda_ethereum_consensus/lib/lambda_ethereum_consensus/state_transition/misc.ex

Line 135 in e01f275

{:ok, index} = compute_shuffled_index(rem(i, total), total, seed)

lambda_ethereum_consensus/lib/lambda_ethereum_consensus/state_transition/accessors.ex

Line 88 in 3da8c9e

|> Misc.compute_shuffled_index(active_validator_count, seed) do

Acceptance criteria:

• There should be some type of benchmark that shows the improvement of performance from using one function vs the other