Perf/parallelize storage commit

[asdacap]

• Parallelize storage commit by committing multiple storage at the same time.
• This need an interface change due to the locking of the dirty nodes.
  • The interface change also remove the need to know if its a state or storage trie and passing of block number to commit.
• Also reduce the threshold at which it start parallel commit.
• This improve block processing time by 5%.
• Graph is before/after/before/after.
  
• Note that before, the storage time include the time it is waiting for a lock while memory pruning is happening. The time just for storage is the storage commit time minus pruning time. The times are all prometheus's rate.

Types of changes

What types of changes does your code introduce?

• Optimization

Testing

Requires testing

• Yes
• No

If yes, did you write tests?

• Yes
• No

Notes on testing

• Mainnet catchup
• Mainnet sync
• Mainnet archive node
• Mainnet forward sync.

[LukaszRozmej]

Overall great, but due to large refactor I think code can be made more readable and there are quite a few possibilities to do that.

It is now confusing what is the difference, usage and purpose of TrieStore.GetTrieStore and the difference between ITrieStore.BeginBlockCommit and IScopedTrieStore.BeginCommit. We either need better naming, or some documenting comments or both.

Also adjust doc comments where method parameters changed.