Improve processEffectiveBalanceUpdates

[twoeths]

Motivation

• It takes so long to process epoch on mainnet (5.4s in my environment)
• I added some log, it showed that it updates around 75000 effective balance of validators per epoch

Description

• Since we have MAX_EFFECTIVE_BALANCE, it turns out that we waste our time to update effective balance for most/all of validators
• We should NOT update if we see that the effective balance is too small to balance, but it's already MAX_EFFECTIVE_BALANCE. New log shows that we have to update 0 effective balance.
• Epoch transition is reduced from 5.4s to ~650ms

Closes #2901

Steps to test or reproduce

• Sync mainnet

[codeclimate]

Code Climate has analyzed commit aa53fed and detected 0 issues on this pull request.

View more on Code Climate.

[github-actions]

Performance Report

✔️ no performance regression detected

Full benchmark results

Benchmark suite	Current: 88e23bc	Previous: 940404e	Ratio
getCommitteeAssignments - req 1000 vs - 250000 vc	8.2858 ms/op	8.2731 ms/op	1.00
epoch altair - 250000 vs - 7PWei - processJustificationAndFinalization	222.12 us/op	192.44 us/op	1.15
epoch altair - 250000 vs - 7PWei - processInactivityUpdates	2.3113 s/op	2.0562 s/op	1.12
epoch altair - 250000 vs - 7PWei - processRewardsAndPenalties	847.13 ms/op	839.28 ms/op	1.01
epoch altair - 250000 vs - 7PWei - processRegistryUpdates	16.488 us/op	16.675 us/op	0.99
epoch altair - 250000 vs - 7PWei - processSlashings	88.551 us/op	96.477 us/op	0.92
epoch altair - 250000 vs - 7PWei - processEffectiveBalanceUpdates	120.37 ms/op	98.628 ms/op	1.22
epoch altair - 250000 vs - 7PWei - processParticipationFlagUpdates	250.83 ms/op	238.27 ms/op	1.05
epoch altair - 250000 vs - 7PWei - prepareEpochProcessState	529.20 ms/op	501.87 ms/op	1.05
Process block - 250000 vs - 7PWei - with 0 validator exit	340.88 us/op	346.07 us/op	0.99
Process block - 250000 vs - 7PWei - with 1 validator exit	41.578 ms/op	38.919 ms/op	1.07
Process block - 250000 vs - 7PWei - with 16 validator exits	29.300 ms/op	21.255 ms/op	1.38
epoch phase0 - 250000 vs - 7PWei - processJustificationAndFinalization	78.186 us/op	90.898 us/op	0.86
epoch phase0 - 250000 vs - 7PWei - processRewardsAndPenalties	424.64 ms/op	470.24 ms/op	0.90
epoch phase0 - 250000 vs - 7PWei - processRegistryUpdates	11.734 us/op	18.627 us/op	0.63
epoch phase0 - 250000 vs - 7PWei - processSlashings	42.152 us/op	34.352 us/op	1.23
epoch phase0 - 250000 vs - 7PWei - processFinalUpdates	50.700 ms/op	39.521 ms/op	1.28
epoch phase0 - 250000 vs - 7PWei - prepareEpochProcessState	856.01 ms/op	825.18 ms/op	1.04
getAttestationDeltas - 250000 vs - 7PWei	87.323 ms/op	88.300 ms/op	0.99
processSlots - 250000 vs - 7PWei - 32 empty slots	5.7464 s/op	5.5973 s/op	1.03
shuffle list - 16384 els	1.6772 ms/op	2.5192 ms/op	0.67
shuffle list - 250000 els	23.935 ms/op	35.786 ms/op	0.67
getPubkeys - persistent - req 1000 vs - 250000 vc	17.274 us/op	19.053 us/op	0.91
BLS verify - blst-native	1.8566 ms/op	1.8660 ms/op	0.99
BLS verifyMultipleSignatures 3 - blst-native	3.8087 ms/op	3.8324 ms/op	0.99
BLS verifyMultipleSignatures 8 - blst-native	8.2063 ms/op	8.2275 ms/op	1.00
BLS verifyMultipleSignatures 32 - blst-native	29.834 ms/op	29.836 ms/op	1.00
BLS aggregatePubkeys 32 - blst-native	40.419 us/op	41.447 us/op	0.98
BLS aggregatePubkeys 128 - blst-native	155.26 us/op	155.46 us/op	1.00
getAttestationsForBlock	140.15 ms/op	126.80 ms/op	1.11
validate gossip signedAggregateAndProof - struct	4.5169 ms/op	4.5062 ms/op	1.00
validate gossip signedAggregateAndProof - treeBacked	4.4688 ms/op	4.4608 ms/op	1.00
validate gossip attestation - struct	2.0976 ms/op	2.0976 ms/op	1.00
validate gossip attestation - treeBacked	2.1331 ms/op	2.1298 ms/op	1.00

by benchmarkbot/action