Benchmark Block Processing Time at Scale

[terencechain]

Background:
Refer to this issue which was discussed in the 3rd eth2.0 implementers call

Goal:
Run some time analysis of block processing for the beacon node. This involves validating both the block and attestations. The following parameters were suggested:
10M ETH deposit, 300k validators, 64 slots
5000 validators per slot
1000 shards across 64 slots which is 16 shards per slot
Each block estimates to 16 attestations with ~300 validators each.

Results from other clients:
Lighthouse (Rust):

BenchmarkParams { total_validators: 312500, cycle_length: 64, shard_count: 1024, shards_per_slot: 16, validators_per_shard: 305 }
test state::block::validation::benches::bench_block_validation ... bench: 345,134,640 ns/iter (+/- 33,842,738)

That's 0.34 sec, Paul said he was able to get it down to 0.11 sec by optimizing the signature library, and 0.062s with concurrency

[terencechain]

#258 is blocking this effort

[djrtwo]

Paul got this test case down to 0.062s with concurrency

[paulhauner]

FYI, these results were on my laptop: a Lenovo X1 Carbon 5th Gen with an Intel i5-7300U @ 2.60GHz running Arch Linux with a low CPU load at the time.

[rauljordan]

We can close this as it is handled automatically by the state transition e2e tests done in the YAML config such as in #1221.