When running the import, currently blocks are loaded in batches into a
`seq` then passed to the importer as such.
In reality, blocks are still processed one by one, so the batching does
not offer any performance advantage. It does however require that the
client wastes memory, up to several GB, on the block sequence while
they're waiting to be processed.
This PR introduces a persister that accepts these potentially large
blocks one by one and at the same time removes a number of redundant /
unnecessary copies, assignments and resets that were slowing down the
import process in general.
The forking facility has been replaced by ForkedChain - frames and
layers are two other mechanisms that mostly do the same thing at the
aristo level, without quite providing the functionality FC needs - this
cleanup will make that integration easier.
The current getCanonicalHead of core db should not be confused with ForkedChain.latestHeader.
Therefore we need to use getCanonicalHead to restricted case only, e.g. initializing ForkedChain.
* Metrics cosmetics
* Better naming for error threshold constants
* Treating header/body process error different from response errors
why:
Error handling becomes active not until some consecutive failures
appear. As both types of errors may interleave (i.g. no response
errors) the counter reset for one type might affect the other.
By doing it wrong, a peer might send repeatedly a bogus block so
locking in the syncer in an endless loop.
In block processing, depending on the complexity of a transaction and
hotness of caches etc, signature checking can actually make up the
majority of time needed to process a transaction (60% observed in some
randomly sampled block ranges).
Fortunately, this is a task that trivially can be offloaded to a task
pool similar to how nimbus-eth2 does it.
This PR introduces taskpools in the most simple way possible, by
performing signature checking concurrently with other TX processing,
assigning a taskpool task per TX effectively.
With this little trick, we're in gigagas land 🎉 on my laptop!
```
INF 2024-12-10 21:05:35.170+01:00 Imported blocks
blockNumber=3874817 b... mgps=1222.707 ...
```
Tests don't use the taskpool for now because it needs manual cleanup and
we don't have a good mechanism in place. Future PR:s should address this
by creating a common shutdown sequence that also closes and cleans up
other resources like the DB.
Co-authored-by: andri lim <jangko128@gmail.com>
* Cosmetics, add some metrics updates to smoothen curves
why:
Progress downloading blocks was just a jump from none to full
* Reclassifying some syncer gossip from TRC to DBG
why:
Might help debugging without full trace logs
Now that branches are small, we can add a branch cache that fits more
verticies in memory by only storing the branch portion (16 bytes) of the
VertexRef (136 bytes).
Where the original vertex cache hovers around a hit rate of ~60:ish,
this branch cache reaches >90% hit rate instead around block 20M which
gives a nice boost to processing.
A downside of this approach is that a new VertexRef must be allocated
for every cache hit instead of reusing an existing instance - this
causes some GC overhead that needs to be addressed.
Nice 15% improvement nonetheless, can't complain!
```
blocks: 19630784, baseline: 161h18m38s, contender: 136h23m23s
Time (total): -24h55m14s, -15.45%
```
Since data is ordered by VertexID on disk, with this simple trick we can
make much better use of the various rocksdb caches.
Computing the state root of the full mainnet state is down to 4 hours
(from 9) on my laptop.
A bit unexpectedly, nibble handling shows up in the profiler mainly
because the current impl is tuned towards slicing while the most common
operation is prefix comparison - since the code is simple, might has
well get rid of some of the excess fat by always aliging the nibbles to
the byte buffer.
* Cosmetics
* Must not async wait inside termination `for` loop
why:
Async-waiting inside a `for` loop will switch to temination process
which uncontrollably will modify the for-loop data base.
* Avoid `waitFor` in scheduler termination code
why:
Is reserved for main loop
* `shouldPrepareTracer` always true
* simple `pop` should not copy value (reading the memory shows up in a
profiler)
* continuation code simplified
* remove some unnecessary EH
* Re-org internal descriptor `CanonicalDesc` as `PivotArc`
why:
Despite its name, `CanonicalDesc` contained a cursor arc (or leg) from
the base tree with a designated block (or Header) on its arc members
(aka blocks.) The type is used more generally than only for s block on
the canonical cursor.
Also, the `PivotArc` provides some more fields for caching intermediate
data. This simplifies managing extra arguments for some functions.
* Remove cruft
details:
No need to find cursor arc if it is given as function argument.
* Rename prototype variables `head: PivotArc` to `pvarc`
why:
Better reading
* Function and code massage, adjust names
details:
Avoid the syllable `canonical` in function names that do not strictly
apply to the canonical chain. So renaming
* findCanonicalHead() => findCursorArc()
* canonicalChain() => findHeader()
* trimCanonicalChain() => trimCursorArc()
* Combine `updateBase()` function-args into single `PivotArgs` object
why:
Will generalise action for more complex scenarios in future.
* update `calculateNewBase()` return code type => `PivotArc`
why:
So it can directly be used as argument into `updateBase()`
* Update `calculateNewBase()` for target on parent arc
* Update unit tests
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
* Kludge: fix `eip4844` import in `validate`
why:
Importing `validate` needs `blscurve` here or with the importing module.
* Separate out `FC` descriptor iinto separate file
why:
Needed for external descriptor access (e.g. for debugging)
* Debugging toolkit for `FC`
* Verify chain descriptor after changing state
* Cosmetics, update log and exception messages
* Update `FC` base tree updater `updateBase()`
why:
Correct `forkJunction` of canonical cursor head record. When moving
the `base`, this field would be below `base` unless updated.
* Fix `FC` chain selector `findCanonicalHead()`
why:
Given a sample ref `hash` the function searched for the unique chain
containing the block header referenced by `hash`.
Unfortunately, when searching down the ancestry lineage, the function
did not necessarily stop an the end of the sub-chain. Rather it
continued with the parent chain without noticing. So returning the
wrong result.
* When calculating new a base it must reside on cursor arc (or leg.)
why:
The finalised block argument (that will eventually be the new base)
might be moved further down the cursor arc if it is too close to the
cursor head (typically smaller than 128 blocks.)
So the finalised block selection is shifted down he cursor arc. And
it might happen that the cursor arc itself is too small and one would
end up at a parent cursor arc. This is rejected.
* Not starting a new cursor arc with a block already on another arc
why:
This leads to an inconsistent set of cursor arcs which are supposed to
be mutually disjunct.
* Tighten condition: A block that is not on the base tree must be on the DB
* One less TODO item
The EVM stack is a hot spot in EVM execution and we end up paying a nim
seq tax in several ways, adding up to ~5% of execution time:
* on initial allocation, all bytes get zeroed - this means we have to
choose between allocating a full stack or just a partial one and then
growing it
* pushing and popping introduce additional zeroing
* reallocations on growth copy + zero - expensive again!
* redundant range checking on every operation reducing inlining etc
Here a custom stack using C memory is instroduced:
* no zeroing on allocation
* full stack allocated on EVM startup -> no reallocation during
execution
* fast push/pop - no zeroing again
* 32-byte alignment - this makes it easier for the compiler to use
vector instructions
* no stack allocated for precompiles (these never use it anyway)
Of course, this change also means we have to manage memory manually -
for the EVM, this turns out to be not too bad because we already manage
database transactions the same way (they have to be freed "manually") so
we can simply latch on to this mechanism.
While we're at it, this PR also skips database lookup for known
precompiles by resolving such addresses earlier.
why:
The `base` block is ancestor to all blocks of the base tree bust stays
outside the tree.
Some fringe condition uses an opportunistic fix when the `cursor` is not in
the base tree, which is legit if `cursor != base`.
* Inline gas cost/instruction fetching
These make up 5:ish % of EVM execution time - even though they're
trivial they end up not being inlined - this little change gives a
practically free perf boost ;)
Also unify the style of creating the output to `setLen`..
* avoid a few more unnecessary seq allocations
* Ignore `FC` overlapping blocks and the ones <= `base`
why:
Due to concurrently running `importBlock()` by `newPayload` RPC
requests the `FC` module layout might differ when re-visiting for
importing blocks.
* Update logging and docu
details:
Reduce some logging noise
Clarify activating/suspending syncer in log messages
`updateOk` is obsolete and always set to true - callers should not have
to care about this detail
also take the opportunity to clean up storage root naming
* Log/trace cancellation events in scheduler
* Provide `clear()` functions for explicitly flushing data objects
* Renaming header cache functions
why:
More systematic, all functions start with prefix `dbHeader`
* Remove `danglingParent` from layout
why:
Already provided by header cache
* Remove `couplerHash` and `headHash` from layout
why:
No need to cache, `headHash` is unused and `couplerHash` used typically
once, only.
* Remove `lastLayout` from sync descriptor
why:
No need to compare changes, saving is always triggered after actively
changing the sync layout state
* Early reject unsuitable head + finalised header from CL
why:
The finalised header is only passed by its hash so the header must be
fetched somewhere, e.g. from a peer via eth/xx.
Also, finalised headers earlier than the `base` from `FC` cannot be
handled due to the `Aristo` single state database architecture.
Luckily, on a full node, the complete block history is available so
unsuitable finalised headers are stored there already which is exploited
here to avoid unnecessary network traffic.
* Code cosmetics, remove cruft, prettify logging, remove `final` metrics
detail:
The `final` layout parameter will be deprecated and later removed
* Update/re-calibrate syncer logic documentation
why:
The current implementation sucks if the `FC` module changes the
canonical branch in the middle of completing a header chain (due
to concurrent updates by the `newPayload()` logic.)
* Implement according to re-calibrated syncer docu
details:
The implementation employs the notion of named layout states (see
`SyncLayoutState` in `worker_desc.nim`) which are derived from the
state parameter triple `(C,D,H)` as described in `README.md`.
Jacek Sieka
andri lim
tersec
Jordan Hrycaj
Advaita Saha