remove the scaffolding get_empty_per_epoch_cache() calls and the pointless construction and return/discarding of zero-filled arrays

This commit is contained in:
Dustin Brody 2020-06-09 09:55:28 +02:00 committed by tersec
parent 5582ecc1f4
commit 35f67cec60
1 changed files with 81 additions and 128 deletions

View File

@ -15,19 +15,12 @@
# The entry point is `process_epoch` which is at the bottom of this file.
#
# General notes about the code (TODO):
# * It's inefficient - we quadratically copy, allocate and iterate when there
# are faster options
# * Weird styling - the sections taken from the spec use python styling while
# the others use NEP-1 - helps grepping identifiers in spec
# * We mix procedural and functional styles for no good reason, except that the
# spec does so also.
# * There are likely lots of bugs.
# * For indices, we get a mix of uint64, ValidatorIndex and int - this is currently
# swept under the rug with casts
# * The spec uses uint64 for data types, but functions in the spec often assume
# signed bigint semantics - under- and overflows ensue
# * Sane error handling is missing in most cases (yay, we'll get the chance to
# debate exceptions again!)
# When updating the code, add TODO sections to mark where there are clear
# improvements to be made - other than that, keep things similar to spec for
# now.
@ -72,7 +65,7 @@ func get_total_active_balance*(state: BeaconState, cache: var StateCache): Gwei
# Note: ``get_total_balance`` returns ``EFFECTIVE_BALANCE_INCREMENT`` Gwei
# minimum to avoid divisions by zero.
# TODO refactor get_epoch_per_epoch_cache() not to be, well, empty, so can
# TODO refactor get_empty_per_epoch_cache() not to be, well, empty, so can
# avoid this ever refilling, and raiseAssert, and get rid of var
let
epoch = state.slot.compute_epoch_at_slot
@ -261,10 +254,8 @@ func get_base_reward(state: BeaconState, index: ValidatorIndex,
effective_balance * BASE_REWARD_FACTOR div
integer_squareroot(total_balance) div BASE_REWARDS_PER_EPOCH
func get_proposer_reward(state: BeaconState, attesting_index: ValidatorIndex):
Gwei =
# TODO REMOVEMEREMOVEME
var cache = get_empty_per_epoch_cache()
func get_proposer_reward(state: BeaconState, attesting_index: ValidatorIndex,
cache: var StateCache): Gwei =
let total_balance = get_total_active_balance(state, cache)
get_base_reward(state, attesting_index, total_balance) div PROPOSER_REWARD_QUOTIENT
@ -285,14 +276,14 @@ func get_eligible_validator_indices(state: BeaconState): seq[ValidatorIndex] =
result.add idx.ValidatorIndex
func get_attestation_component_deltas(state: BeaconState,
attestations: seq[PendingAttestation]
): tuple[a: seq[Gwei], b: seq[Gwei]] =
attestations: seq[PendingAttestation],
cache: var StateCache,
): tuple[a: seq[Gwei], b: seq[Gwei]] =
# Helper with shared logic for use by get source, target, and head deltas
# functions
var
rewards = repeat(0'u64, len(state.validators))
penalties = repeat(0'u64, len(state.validators))
cache = get_empty_per_epoch_cache() # REMOVEMEBEFOREMERGE
let
total_balance = get_total_active_balance(state, cache)
unslashed_attesting_indices =
@ -316,31 +307,34 @@ func get_attestation_component_deltas(state: BeaconState,
(rewards, penalties)
# https://github.com/ethereum/eth2.0-specs/blob/v0.12.1/specs/phase0/beacon-chain.md#components-of-attestation-deltas
func get_source_deltas(state: BeaconState): tuple[a: seq[Gwei], b: seq[Gwei]] =
func get_source_deltas(state: BeaconState, cache: var StateCache):
tuple[a: seq[Gwei], b: seq[Gwei]] =
# Return attester micro-rewards/penalties for source-vote for each validator.
let matching_source_attestations =
get_matching_source_attestations(state, get_previous_epoch(state))
get_attestation_component_deltas(state, matching_source_attestations)
get_attestation_component_deltas(state, matching_source_attestations, cache)
func get_target_deltas(state: BeaconState): tuple[a: seq[Gwei], b: seq[Gwei]] =
func get_target_deltas(state: BeaconState, cache: var StateCache):
tuple[a: seq[Gwei], b: seq[Gwei]] =
# Return attester micro-rewards/penalties for target-vote for each validator.
let matching_target_attestations =
get_matching_target_attestations(state, get_previous_epoch(state))
get_attestation_component_deltas(state, matching_target_attestations)
get_attestation_component_deltas(state, matching_target_attestations, cache)
func get_head_deltas(state: BeaconState): tuple[a: seq[Gwei], b: seq[Gwei]] =
func get_head_deltas(state: BeaconState, cache: var StateCache):
tuple[a: seq[Gwei], b: seq[Gwei]] =
# Return attester micro-rewards/penalties for head-vote for each validator.
let matching_head_attestations =
get_matching_head_attestations(state, get_previous_epoch(state))
get_attestation_component_deltas(state, matching_head_attestations)
get_attestation_component_deltas(state, matching_head_attestations, cache)
func get_inclusion_delay_deltas(state: BeaconState): tuple[a: seq[Gwei], b: seq[Gwei]] =
func get_inclusion_delay_deltas(state: BeaconState, cache: var StateCache):
seq[Gwei] =
# Return proposer and inclusion delay micro-rewards/penalties for each validator.
var rewards = repeat(0'u64, len(state.validators))
let matching_source_attestations =
get_matching_source_attestations(state, get_previous_epoch(state))
# TODO some fun big-O's here; see get_attestation_deltas_old() for linear version
var cache = get_empty_per_epoch_cache()
let total_balance = get_total_active_balance(state, cache)
for index in get_unslashed_attesting_indices(state, matching_source_attestations, cache):
# this is bad; it's to be just-correct-enough to check if v0.12.1 beacon chain is working
@ -348,19 +342,20 @@ func get_inclusion_delay_deltas(state: BeaconState): tuple[a: seq[Gwei], b: seq[
let admi = minIndex(mapIt(attestations, it.inclusion_delay))
let attestation = attestations[admi]
rewards[attestation.proposer_index] += get_proposer_reward(state, index)
# TODO remove duplicate calculation of get_proposer_reward()
rewards[attestation.proposer_index] += get_proposer_reward(state, index, cache)
let max_attester_reward =
get_base_reward(state, index, total_balance) - get_proposer_reward(state, index)
get_base_reward(state, index, total_balance) - get_proposer_reward(state, index, cache)
rewards[index] += Gwei(max_attester_reward div attestation.inclusion_delay)
# No penalties associated with inclusion delay
let penalties = repeat(0'u64, len(state.validators))
(rewards, penalties)
# Spec constructs both and returns both; this doesn't
rewards
func get_inactivity_penalty_deltas(state: BeaconState): tuple[a: seq[Gwei], b: seq[Gwei]] =
func get_inactivity_penalty_deltas(state: BeaconState, cache: var StateCache):
seq[Gwei] =
# Return inactivity reward/penalty deltas for each validator.
var penalties = repeat(0'u64, len(state.validators))
var cache = get_empty_per_epoch_cache() # REMOVEME
let total_balance = get_total_active_balance(state, cache) # DO NOT KEEP RECALCULATING THIS
if is_in_inactivity_leak(state):
let
@ -372,133 +367,91 @@ func get_inactivity_penalty_deltas(state: BeaconState): tuple[a: seq[Gwei], b: s
for index in get_eligible_validator_indices(state):
# If validator is performing optimally this cancels all rewards for a neutral balance
let base_reward = get_base_reward(state, index, total_balance)
penalties[index] += Gwei(BASE_REWARDS_PER_EPOCH * base_reward - get_proposer_reward(state, index))
penalties[index] +=
Gwei(BASE_REWARDS_PER_EPOCH * base_reward -
get_proposer_reward(state, index, cache))
if index notin matching_target_attesting_indices:
let effective_balance = state.validators[index].effective_balance
penalties[index] +=
Gwei(effective_balance * get_finality_delay(state) div INACTIVITY_PENALTY_QUOTIENT)
# No rewards associated with inactivity penalties
let rewards = repeat(0'u64, len(state.validators))
(rewards, penalties)
# Spec constructs rewards anyway; this doesn't
penalties
# https://github.com/ethereum/eth2.0-specs/blob/v0.12.1/specs/phase0/beacon-chain.md#get_attestation_deltas
func get_attestation_deltas_new(state: BeaconState, cache: var StateCache): tuple[a: seq[Gwei], b: seq[Gwei]] =
func get_attestation_deltas(state: BeaconState, cache: var StateCache): tuple[a: seq[Gwei], b: seq[Gwei]] =
# Return attestation reward/penalty deltas for each validator.
let
(source_rewards, source_penalties) = get_source_deltas(state)
(target_rewards, target_penalties) = get_target_deltas(state)
(head_rewards, head_penalties) = get_head_deltas(state)
(inclusion_delay_rewards, _) = get_inclusion_delay_deltas(state)
(_, inactivity_penalties) = get_inactivity_penalty_deltas(state)
(source_rewards, source_penalties) = get_source_deltas(state, cache)
(target_rewards, target_penalties) = get_target_deltas(state, cache)
(head_rewards, head_penalties) = get_head_deltas(state, cache)
inclusion_delay_rewards = get_inclusion_delay_deltas(state, cache)
inactivity_penalties = get_inactivity_penalty_deltas(state, cache)
let rewards = mapIt(0 ..< len(state.validators),
source_rewards[it] + target_rewards[it] + head_rewards[it] + inclusion_delay_rewards[it])
source_rewards[it] + target_rewards[it] + head_rewards[it] +
inclusion_delay_rewards[it])
let penalties = mapIt(0 ..< len(state.validators),
source_penalties[it] + target_penalties[it] + head_penalties[it] + inactivity_penalties[it])
source_penalties[it] + target_penalties[it] + head_penalties[it] +
inactivity_penalties[it])
(rewards, penalties)
# https://github.com/ethereum/eth2.0-specs/blob/v0.11.3/specs/phase0/beacon-chain.md#rewards-and-penalties-1
func get_attestation_deltas_old(state: BeaconState, stateCache: var StateCache):
tuple[a: seq[Gwei], b: seq[Gwei]] =
let
previous_epoch = get_previous_epoch(state)
total_balance = get_total_active_balance(state, stateCache)
var
rewards = repeat(0'u64, len(state.validators))
penalties = repeat(0'u64, len(state.validators))
eligible_validator_indices : seq[ValidatorIndex] = @[]
when false:
## This depends on matching_source_attestations being an indexable seq, not a
## set, hash table, etc.
let source_attestation_attesting_indices =
mapIt(
matching_source_attestations,
get_attesting_indices(state, it.data, it.aggregation_bits, stateCache))
for index, v in state.validators:
if is_active_validator(v, previous_epoch) or
(v.slashed and previous_epoch + 1 < v.withdrawable_epoch):
eligible_validator_indices.add index.ValidatorIndex
for index in get_unslashed_attesting_indices(
state, matching_source_attestations, stateCache):
# Translation of attestation = min([...])
doAssert matching_source_attestations.len > 0
# Micro-incentives for matching FFG source, FFG target, and head
let
matching_source_attestations =
get_matching_source_attestations(state, previous_epoch)
matching_target_attestations =
get_matching_target_attestations(state, previous_epoch)
matching_head_attestations =
get_matching_head_attestations(state, previous_epoch)
# Start by filtering the right attestations
var filtered_matching_source_attestations: seq[PendingAttestation]
for attestations in
[matching_source_attestations, matching_target_attestations,
matching_head_attestations]:
let
unslashed_attesting_indices =
get_unslashed_attesting_indices(state, attestations, stateCache)
attesting_balance = get_total_balance(state, unslashed_attesting_indices)
for index in eligible_validator_indices:
if index in unslashed_attesting_indices:
# Factored out from balance totals to avoid uint64 overflow
const increment = EFFECTIVE_BALANCE_INCREMENT
let reward_numerator = get_base_reward(state, index, total_balance) *
(attesting_balance div increment)
rewards[index] += reward_numerator div (total_balance div increment)
else:
penalties[index] += get_base_reward(state, index, total_balance)
for source_attestation_index, a in matching_source_attestations:
if index notin
source_attestation_attesting_indices[source_attestation_index]:
continue
filtered_matching_source_attestations.add a
# Proposer and inclusion delay micro-rewards
## This depends on matching_source_attestations being an indexable seq, not a
## set, hash table, etc.
let source_attestation_attesting_indices =
mapIt(
matching_source_attestations,
get_attesting_indices(state, it.data, it.aggregation_bits, stateCache))
# The first filtered attestation serves as min until we find something
# better
var attestation = filtered_matching_source_attestations[0]
for source_attestation_index, a in filtered_matching_source_attestations:
if a.inclusion_delay < attestation.inclusion_delay:
attestation = a
for index in get_unslashed_attesting_indices(
state, matching_source_attestations, stateCache):
# Translation of attestation = min([...])
doAssert matching_source_attestations.len > 0
let
base_reward = get_base_reward(state, index, total_balance)
proposer_reward = (base_reward div PROPOSER_REWARD_QUOTIENT).Gwei
# Start by filtering the right attestations
var filtered_matching_source_attestations: seq[PendingAttestation]
rewards[attestation.proposer_index.int] += proposer_reward
let max_attester_reward = base_reward - proposer_reward
for source_attestation_index, a in matching_source_attestations:
if index notin
source_attestation_attesting_indices[source_attestation_index]:
continue
filtered_matching_source_attestations.add a
rewards[index] += max_attester_reward div attestation.inclusion_delay
# The first filtered attestation serves as min until we find something
# better
var attestation = filtered_matching_source_attestations[0]
for source_attestation_index, a in filtered_matching_source_attestations:
if a.inclusion_delay < attestation.inclusion_delay:
attestation = a
let
base_reward = get_base_reward(state, index, total_balance)
proposer_reward = (base_reward div PROPOSER_REWARD_QUOTIENT).Gwei
rewards[attestation.proposer_index.int] += proposer_reward
let max_attester_reward = base_reward - proposer_reward
rewards[index] += max_attester_reward div attestation.inclusion_delay
# Inactivity penalty
let finality_delay = previous_epoch - state.finalized_checkpoint.epoch
if finality_delay > MIN_EPOCHS_TO_INACTIVITY_PENALTY:
let matching_target_attesting_indices =
get_unslashed_attesting_indices(
state, matching_target_attestations, stateCache)
for index in eligible_validator_indices:
penalties[index] +=
BASE_REWARDS_PER_EPOCH.uint64 * get_base_reward(state, index, total_balance)
if index notin matching_target_attesting_indices:
# Inactivity penalty
let finality_delay = previous_epoch - state.finalized_checkpoint.epoch
if finality_delay > MIN_EPOCHS_TO_INACTIVITY_PENALTY:
let matching_target_attesting_indices =
get_unslashed_attesting_indices(
state, matching_target_attestations, stateCache)
for index in eligible_validator_indices:
penalties[index] +=
state.validators[index].effective_balance *
finality_delay div INACTIVITY_PENALTY_QUOTIENT
(rewards, penalties)
func get_attestation_deltas(state: BeaconState, stateCache: var StateCache):
tuple[a: seq[Gwei], b: seq[Gwei]] {.nbench.}=
get_attestation_deltas_new(state, stateCache)
BASE_REWARDS_PER_EPOCH.uint64 * get_base_reward(state, index, total_balance)
if index notin matching_target_attesting_indices:
penalties[index] +=
state.validators[index].effective_balance *
finality_delay div INACTIVITY_PENALTY_QUOTIENT
# https://github.com/ethereum/eth2.0-specs/blob/v0.11.3/specs/phase0/beacon-chain.md#rewards-and-penalties-1
func process_rewards_and_penalties(