nimbus-eth1/nimbus/db/aristo/aristo_serialise.nim
Jacek Sieka f034af422a
Pre-allocate vids for branches (#2882)
Each branch node may have up to 16 sub-items - currently, these are
given VertexID based when they are first needed leading to a
mostly-random order of vertexid for each subitem.

Here, we pre-allocate all 16 vertex ids such that when a branch subitem
is filled, it already has a vertexid waiting for it. This brings several
important benefits:

* subitems are sorted and "close" in their id sequencing - this means
that when rocksdb stores them, they are likely to end up in the same
data block thus improving read efficiency
* because the ids are consequtive, we can store just the starting id and
a bitmap representing which subitems are in use - this reduces disk
space usage for branches allowing more of them fit into a single disk
read, further improving disk read and caching performance - disk usage
at block 18M is down from 84 to 78gb!
* the in-memory footprint of VertexRef reduced allowing more instances
to fit into caches and less memory to be used overall.

Because of the increased locality of reference, it turns out that we no
longer need to iterate over the entire database to efficiently generate
the hash key database because the normal computation is now faster -
this significantly benefits "live" chain processing as well where each
dirtied key must be accompanied by a read of all branch subitems next to
it - most of the performance benefit in this branch comes from this
locality-of-reference improvement.

On a sample resync, there's already ~20% improvement with later blocks
seeing increasing benefit (because the trie is deeper in later blocks
leading to more benefit from branch read perf improvements)

```
blocks: 18729664, baseline: 190h43m49s, contender: 153h59m0s
Time (total): -36h44m48s, -19.27%
```

Note: clients need to be resynced as the PR changes the on-disk format

R.I.P. little bloom filter - your life in the repo was short but
valuable
2024-12-04 11:42:04 +01:00

152 lines
4.4 KiB
Nim

# nimbus-eth1
# Copyright (c) 2023-2024 Status Research & Development GmbH
# Licensed under either of
# * Apache License, version 2.0, ([LICENSE-APACHE](LICENSE-APACHE) or
# http://www.apache.org/licenses/LICENSE-2.0)
# * MIT license ([LICENSE-MIT](LICENSE-MIT) or
# http://opensource.org/licenses/MIT)
# at your option. This file may not be copied, modified, or distributed
# except according to those terms.
{.push raises: [].}
import
eth/[common, rlp],
results,
"."/[aristo_constants, aristo_desc, aristo_compute]
type
ResolveVidFn = proc(
vid: VertexID;
): Result[HashKey,AristoError]
{.gcsafe, raises: [].}
## Resolve storage root vertex ID
# ------------------------------------------------------------------------------
# Private helper
# ------------------------------------------------------------------------------
proc serialise(
pyl: LeafPayload;
getKey: ResolveVidFn;
): Result[seq[byte],(VertexID,AristoError)] =
## Encode the data payload of the argument `pyl` as RLP `seq[byte]` if it is
## of account type, otherwise pass the data as is.
##
case pyl.pType:
of AccountData:
let key = block:
if pyl.stoID.isValid:
pyl.stoID.vid.getKey.valueOr:
let w = (pyl.stoID.vid, error)
return err(w)
else:
VOID_HASH_KEY
ok rlp.encode Account(
nonce: pyl.account.nonce,
balance: pyl.account.balance,
storageRoot: key.to(Hash32),
codeHash: pyl.account.codeHash)
of StoData:
ok rlp.encode pyl.stoData
# ------------------------------------------------------------------------------
# Public RLP transcoder mixins
# ------------------------------------------------------------------------------
proc to*(node: NodeRef; T: type seq[seq[byte]]): T =
## Convert the argument pait `w` to a single or a double item list item of
## `<rlp-encoded-node>` type entries. Only in case of a combined extension
## and branch vertex argument, there will be a double item list result.
##
case node.vtx.vType:
of Branch:
# Do branch node
var wr = initRlpWriter()
wr.startList(17)
for n in 0..15:
wr.append node.key[n]
wr.append EmptyBlob
let brData = wr.finish()
if 0 < node.vtx.pfx.len:
# Prefix branch by embedded extension node
let brHash = brData.digestTo(HashKey)
var wrx = initRlpWriter()
wrx.startList(2)
wrx.append node.vtx.pfx.toHexPrefix(isleaf = false).data()
wrx.append brHash
result.add wrx.finish()
result.add brData
else:
# Do for pure branch node
result.add brData
of Leaf:
proc getKey0(
vid: VertexID;
): Result[HashKey,AristoError]
{.gcsafe, raises: [].} =
ok(node.key[0]) # always succeeds
var wr = initRlpWriter()
wr.startList(2)
wr.append node.vtx.pfx.toHexPrefix(isleaf = true).data()
wr.append node.vtx.lData.serialise(getKey0).value
result.add (wr.finish())
proc digestTo*(node: NodeRef; T: type HashKey): T =
## Convert the argument `node` to the corresponding Merkle hash key. Note
## that a `Dummy` node is encoded as as a `Leaf`.
##
var wr = initRlpWriter()
case node.vtx.vType:
of Branch:
# Do branch node
wr.startList(17)
for n in 0..15:
wr.append node.key[n]
wr.append EmptyBlob
# Do for embedded extension node
if 0 < node.vtx.pfx.len:
let brHash = wr.finish().digestTo(HashKey)
wr = initRlpWriter()
wr.startList(2)
wr.append node.vtx.pfx.toHexPrefix(isleaf = false).data()
wr.append brHash
of Leaf:
proc getKey0(
vid: VertexID;
): Result[HashKey,AristoError]
{.gcsafe, raises: [].} =
ok(node.key[0]) # always succeeds
wr.startList(2)
wr.append node.vtx.pfx.toHexPrefix(isleaf = true).data()
wr.append node.vtx.lData.serialise(getKey0).value
wr.finish().digestTo(HashKey)
proc serialise*(
db: AristoDbRef;
root: VertexID;
pyl: LeafPayload;
): Result[seq[byte],(VertexID,AristoError)] =
## Encode the data payload of the argument `pyl` as RLP `seq[byte]` if it is
## of account type, otherwise pass the data as is.
##
proc getKey(vid: VertexID): Result[HashKey,AristoError] =
ok (?db.computeKey((root, vid)))
pyl.serialise getKey
# ------------------------------------------------------------------------------
# End
# ------------------------------------------------------------------------------