Improve block_sim speed with 50k+ validators

beacon_chain.nimble

@@ -76,5 +76,5 @@ task test, "Run all tests":
   buildAndRunBinary "all_fixtures_require_ssz", "tests/official/", "-d:chronicles_log_level=TRACE -d:const_preset=mainnet"
 
   # State and block sims; getting to 4th epoch triggers consensus checks
-  buildAndRunBinary "state_sim", "research/", "-d:const_preset=mainnet", "--validators=2000 --slots=128"
-  buildAndRunBinary "block_sim", "research/", "-d:const_preset=mainnet", "--validators=2000 --slots=128"
+  buildAndRunBinary "state_sim", "research/", "-d:const_preset=mainnet", "--validators=3000 --slots=128"
+  buildAndRunBinary "block_sim", "research/", "-d:const_preset=mainnet", "--validators=3000 --slots=128"

beacon_chain/attestation_pool.nim

@@ -9,7 +9,7 @@
 
 import
   # Standard libraries
-  deques, sequtils, tables, options, algorithm,
+  deques, sequtils, tables, options,
   # Status libraries
   chronicles, stew/[byteutils], json_serialization/std/sets,
   # Internal
@@ -274,8 +274,6 @@ proc addResolved(pool: var AttestationPool, blck: BlockRef, attestation: Attesta
           #      sets by virtue of not overlapping with some other attestation
           #      and therefore being useful after all?
           trace "Ignoring subset attestation",
-            existingParticipants = get_attesting_indices_seq(
-              state, a.data, v.aggregation_bits, cache),
             newParticipants = participants,
             cat = "filtering"
           found = true
@@ -286,10 +284,6 @@ proc addResolved(pool: var AttestationPool, blck: BlockRef, attestation: Attesta
         # can now be removed per same logic as above
 
         trace "Removing subset attestations",
-          existingParticipants = a.validations.filterIt(
-            it.aggregation_bits.isSubsetOf(validation.aggregation_bits)
-          ).mapIt(get_attesting_indices_seq(
-            state, a.data, it.aggregation_bits, cache)),
           newParticipants = participants,
           cat = "pruning"
 

beacon_chain/beacon_node.nim

@@ -417,7 +417,7 @@ proc onSlotStart(node: BeaconNode, lastSlot, scheduledSlot: Slot) {.gcsafe, asyn
       epoch = scheduledSlot.compute_epoch_at_slot(),
       slot = scheduledSlot
 
-    # Brute-force, but ensure it's reliably enough to run in CI.
+    # Brute-force, but ensure it's reliable enough to run in CI.
     quit(0)
 
   if not wallSlot.afterGenesis or (wallSlot.slot < lastSlot):

beacon_chain/spec/datatypes.nim

@@ -460,7 +460,6 @@ type
   StateCache* = object
     shuffled_active_validator_indices*:
       Table[Epoch, seq[ValidatorIndex]]
-    committee_count_cache*: Table[Epoch, uint64]
     beacon_proposer_indices*: Table[Slot, Option[ValidatorIndex]]
 
 func shortValidatorKey*(state: BeaconState, validatorIdx: int): string =

beacon_chain/spec/state_transition_epoch.nim

@@ -22,7 +22,7 @@
 {.push raises: [Defect].}
 
 import
-  math, sequtils, tables,
+  math, sequtils, tables, algorithm,
   stew/[bitops2], chronicles, json_serialization/std/sets,
   metrics, ../extras, ../ssz/merkleization,
   beaconstate, crypto, datatypes, digest, helpers, validator,
@@ -330,41 +330,41 @@ func get_inclusion_delay_deltas(
     state: BeaconState, total_balance: Gwei, 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
+  var
+    rewards = repeat(0'u64, len(state.validators))
     matching_source_attestations =
       get_matching_source_attestations(state, get_previous_epoch(state))
-    source_attestation_attesting_indices = mapIt(
-      matching_source_attestations,
-      get_attesting_indices(state, it.data, it.aggregation_bits, cache))
-
-  for index in get_unslashed_attesting_indices(state, matching_source_attestations, cache):
-    # Translation of attestation = min([...])
-    # Start by filtering the right attestations
-    var filtered_matching_source_attestations: seq[PendingAttestation]
 
-    for source_attestation_index, a in matching_source_attestations:
-      if index notin
+  # Translation of attestation = min([...])
+  # The spec (pseudo)code defines this in terms of Python's min(), which per
+  # https://docs.python.org/3/library/functions.html#min:
+  # If multiple items are minimal, the function returns the first one
+  # encountered.
+  # Therefore, this approach depends on Nim's default sort being stable, per
+  # https://nim-lang.org/docs/algorithm.html#sort,openArray[T],proc(T,T) via
+  # "The sorting is guaranteed to be stable and the worst case is guaranteed
+  # to be O(n log n)."
+  matching_source_attestations.sort do (x, y: PendingAttestation) -> int:
+    cmp(x.inclusion_delay, y.inclusion_delay)
+
+  # Order/indices in source_attestation_attesting_indices matches sorted order
+  let source_attestation_attesting_indices = mapIt(
+    matching_source_attestations,
+    get_attesting_indices(state, it.data, it.aggregation_bits, cache))
+
+  for index in get_unslashed_attesting_indices(
+      state, matching_source_attestations, cache):
+    for source_attestation_index, attestation in matching_source_attestations:
+      if index in
           source_attestation_attesting_indices[source_attestation_index]:
-        continue
-      filtered_matching_source_attestations.add a
-
-    if filtered_matching_source_attestations.len == 0:
-      continue
-
-    # 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
-
-    rewards[attestation.proposer_index] +=
-      get_proposer_reward(state, index, total_balance)
-    let max_attester_reward =
-      get_base_reward(state, index, total_balance) -
-        get_proposer_reward(state, index, total_balance)
-    rewards[index] += Gwei(max_attester_reward div attestation.inclusion_delay)
+        rewards[attestation.proposer_index] +=
+          get_proposer_reward(state, index, total_balance)
+        let max_attester_reward =
+          get_base_reward(state, index, total_balance) -
+            get_proposer_reward(state, index, total_balance)
+        rewards[index] +=
+          Gwei(max_attester_reward div attestation.inclusion_delay)
+        break
 
   # No penalties associated with inclusion delay
   # Spec constructs both and returns both; this doesn't

beacon_chain/spec/validator.nim

@@ -9,7 +9,7 @@
 {.push raises: [Defect].}
 
 import
-  algorithm, options, sequtils, math, tables,
+  algorithm, options, sequtils, math, tables, sets,
   ./datatypes, ./digest, ./helpers
 
 # https://github.com/ethereum/eth2.0-specs/blob/v0.12.1/specs/phase0/beacon-chain.md#compute_shuffled_index
@@ -140,18 +140,15 @@ func get_beacon_committee*(
     cache.shuffled_active_validator_indices[epoch] =
       get_shuffled_active_validator_indices(state, epoch)
 
-  # Constant throughout an epoch
-  if epoch notin cache.committee_count_cache:
-    cache.committee_count_cache[epoch] =
-      get_committee_count_at_slot(state, slot)
-
   try:
+    let committee_count = get_committee_count_at_slot(
+      cache.shuffled_active_validator_indices[epoch].len)
     compute_committee(
       cache.shuffled_active_validator_indices[epoch],
       get_seed(state, epoch, DOMAIN_BEACON_ATTESTER),
-      (slot mod SLOTS_PER_EPOCH) * cache.committee_count_cache[epoch] +
+      (slot mod SLOTS_PER_EPOCH) * committee_count +
         index.uint64,
-      cache.committee_count_cache[epoch] * SLOTS_PER_EPOCH
+      committee_count * SLOTS_PER_EPOCH
     )
   except KeyError:
     raiseAssert "values are added to cache before using them"
@@ -160,7 +157,6 @@ func get_beacon_committee*(
 func get_empty_per_epoch_cache*(): StateCache =
   result.shuffled_active_validator_indices =
     initTable[Epoch, seq[ValidatorIndex]]()
-  result.committee_count_cache = initTable[Epoch, uint64]()
 
 # https://github.com/ethereum/eth2.0-specs/blob/v0.12.1/specs/phase0/beacon-chain.md#compute_proposer_index
 func compute_proposer_index(state: BeaconState, indices: seq[ValidatorIndex],
@@ -234,14 +230,18 @@ func get_beacon_proposer_index*(state: BeaconState, cache: var StateCache):
 
 # https://github.com/ethereum/eth2.0-specs/blob/v0.12.1/specs/phase0/validator.md#validator-assignments
 func get_committee_assignment*(
-    state: BeaconState, epoch: Epoch, validator_index: ValidatorIndex):
+    state: BeaconState, epoch: Epoch,
+    validator_indices: HashSet[ValidatorIndex]):
     Option[tuple[a: seq[ValidatorIndex], b: CommitteeIndex, c: Slot]] {.used.} =
   # Return the committee assignment in the ``epoch`` for ``validator_index``.
   # ``assignment`` returned is a tuple of the following form:
   #     * ``assignment[0]`` is the list of validators in the committee
   #     * ``assignment[1]`` is the index to which the committee is assigned
   #     * ``assignment[2]`` is the slot at which the committee is assigned
   # Return None if no assignment.
+  #
+  # Slightly adapted from spec version to support multiple validator indices,
+  # since each beacon_node supports many validators.
   let next_epoch = get_current_epoch(state) + 1
   doAssert epoch <= next_epoch
 
@@ -251,9 +251,8 @@ func get_committee_assignment*(
   for slot in start_slot ..< start_slot + SLOTS_PER_EPOCH:
     for index in 0 ..< get_committee_count_at_slot(state, slot):
       let idx = index.CommitteeIndex
-      let committee =
-        get_beacon_committee(state, slot, idx, cache)
-      if validator_index in committee:
+      let committee = get_beacon_committee(state, slot, idx, cache)
+      if not disjoint(validator_indices, toHashSet(committee)):
         return some((committee, idx, slot))
   none(tuple[a: seq[ValidatorIndex], b: CommitteeIndex, c: Slot])
 

beacon_chain/validator_api.nim

@@ -7,7 +7,7 @@
 
 import
   # Standard library
-  tables, strutils, parseutils, sequtils,
+  tables, strutils, parseutils, sequtils, sets,
 
   # Nimble packages
   stew/[byteutils, objects],
@@ -356,7 +356,8 @@ proc installValidatorApiHandlers*(rpcServer: RpcServer, node: BeaconNode) =
         let idx = state.validators.asSeq.findIt(it.pubKey == pubkey)
         if idx == -1:
           continue
-        let ca = state.get_committee_assignment(epoch, idx.ValidatorIndex)
+        let ca = state.get_committee_assignment(
+          epoch, toHashSet([idx.ValidatorIndex]))
         if ca.isSome:
           result.add((public_key: pubkey,
                       committee_index: ca.get.b,