Gas optimizations for deposit process

[vshvsh]

Currently, the submit function, compiled with 200 optimizer iterations, consumes around 577000 gas plus 107000 gas for each registered validator, so the default limit is set by deploy scripts to 16 validators to make the submit transaction occupy no more than 20% of a block. See this file for the related calculations; you can run them with yarn estimate-deposit-loop-gas.

That is on average 130k+ gas per validation key in a batch of 16, about twice as much as a simple deposit. We might find a way to make this number lower.

Note: Switching to the official Deposit contract #88 will impact the numbers mentioned above

[ujenjt]

I've set up an experiment and measure gas-consuming by different deposit transactions. Initially, I've added 16 Staking Providers (SP) and then measure how much gas consumes the call of pool.deposit with 33 ETH, 65ETH and 97ETH. Currently, I am using the mock instead of the real deposit contract. I've extracted the most gas-consuming stages from the deposit transactions. See the table below.

Description	33 ETH	65 ETH	97 ETH
tx total gas	862.6k	1161k	1459k
minting StETH	100k	100k	100k
load all SPs from storage	300k	300k	300k
load signing keys from storage	22k	45k	68k
perform stake to mock deposit	260k	516k	762k
save updated SPs to storage	30k	38k	48k
the rest of the code	150k	162k	181k

I've attached a row output of the profiling for that three transactions.

16sp_33eth_deposit.txt
16SP_65eth_deposit.txt
16SP_97eth_deposit.txt

[ujenjt]

Fixed in #107 and #225