chore(db): microbenchmarking for table serialization and db insertion (#513)

Co-authored-by: Georgios Konstantopoulos <me@gakonst.com>

.github/scripts/compare_iai.sh

@@ -0,0 +1,6 @@
+#!/bin/bash
+
+# This script should be run on the main branch, after running the iai benchmarks on the target branch.
+
+# If the main branch has a better iai performance, exits in error. It ignores L2 differences, since they seem hard to stabilize across runs.
+cargo bench --package reth-db --bench iai | tee /dev/tty | awk '/((L1)|(Ins)|(RAM)|(Est))+.*\(\+[1-9]+[0-9]*\..*%\)/{f=1} END{exit f}'

.github/workflows/ci.yml

@@ -182,3 +182,43 @@ jobs:
       - uses: Swatinem/rust-cache@v2
       - name: Check if documentation builds
         run: RUSTDOCFLAGS="-D warnings" cargo doc --all --no-deps --all-features --document-private-items
+
+  benchmarks:
+    # Pin to `20.04` instead of `ubuntu-latest`, until ubuntu-latest migration is complete
+    # See also <https://github.com/foundry-rs/foundry/issues/3827>
+    runs-on: ubuntu-20.04
+    steps:
+      - name: Install Valgrind
+        run: |
+          sudo apt install valgrind
+
+      - name: Checkout PR sources
+        uses: actions/checkout@v3
+        with:
+          ref: main
+
+      - uses: Swatinem/rust-cache@v1
+        with:
+          cache-on-failure: true
+
+      - name: Generate test-vectors
+        uses: actions-rs/cargo@v1
+        with:
+          command: run
+          args: --bin reth -- test-vectors tables
+
+      - name: Set main baseline
+        uses: actions-rs/cargo@v1
+        with:
+          command: bench
+          args: --package reth-db --bench iai
+
+      - name: Checkout main sources
+        uses: actions/checkout@v3
+        with:
+          clean: false
+
+      - name: Compare PR benchmark
+        shell: 'script -q -e -c "bash {0}"' # required to workaround /dev/tty not being available
+        run: |
+          ./.github/scripts/compare_iai.sh

.gitignore

@@ -15,3 +15,6 @@ target/
 
 # Generated by MacOS
 .DS_Store
+
+# Generated test-vectors for DB
+testdata/micro/db

Cargo.lock

@@ -3970,6 +3970,7 @@ dependencies = [
  "metrics",
  "metrics-exporter-prometheus",
  "metrics-util",
+ "proptest",
  "reth-consensus",
  "reth-db",
  "reth-discv4",
@@ -4052,6 +4053,7 @@ dependencies = [
  "modular-bitfield",
  "page_size",
  "parity-scale-codec",
+ "paste",
  "postcard",
  "proptest",
  "proptest-derive",
@@ -4063,6 +4065,7 @@ dependencies = [
  "reth-primitives",
  "secp256k1 0.24.2",
  "serde",
+ "serde_json",
  "tempfile",
  "test-fuzz",
  "thiserror",

bin/reth/Cargo.toml

@@ -8,7 +8,7 @@ readme = "README.md"
 
 [dependencies]
 # reth
-reth-primitives = { path = "../../crates/primitives" }
+reth-primitives = { path = "../../crates/primitives", features = ["arbitrary"] }
 reth-db = {path = "../../crates/storage/db", features = ["mdbx", "test-utils"] }
 # TODO: Temporary use of the test-utils feature
 reth-provider = { path = "../../crates/storage/provider", features = ["test-utils"] }
@@ -44,6 +44,9 @@ metrics = "0.20.1"
 metrics-exporter-prometheus = { version = "0.11.0", features = ["http-listener"] }
 metrics-util = "0.14.0"
 
+# test vectors generation
+proptest = "1.0"
+
 # misc
 eyre = "0.6.8"
 clap = { version = "4.0", features = ["derive", "cargo"] }

bin/reth/src/cli.rs

@@ -2,7 +2,7 @@
 use crate::{
     db,
     dirs::{LogsDir, PlatformPath},
-    node, p2p, stage, test_eth_chain,
+    node, p2p, stage, test_eth_chain, test_vectors,
 };
 use clap::{ArgAction, Args, Parser, Subcommand};
 use reth_tracing::{
@@ -25,6 +25,7 @@ pub async fn run() -> eyre::Result<()> {
         Commands::Db(command) => command.execute().await,
         Commands::Stage(command) => command.execute().await,
         Commands::P2P(command) => command.execute().await,
+        Commands::TestVectors(command) => command.execute().await,
     }
 }
 
@@ -51,6 +52,9 @@ pub enum Commands {
     /// Run Ethereum blockchain tests
     #[command(name = "test-chain")]
     TestEthChain(test_eth_chain::Command),
+    /// Generate Test Vectors
+    #[command(name = "test-vectors")]
+    TestVectors(test_vectors::Command),
 }
 
 #[derive(Parser)]

bin/reth/src/db/mod.rs

@@ -170,6 +170,7 @@ impl Command {
                     BlockBodies,
                     BlockOmmers,
                     TxHashNumber,
+                    PlainStorageState,
                     PlainAccountState,
                     BlockTransitionIndex,
                     TxTransitionIndex,

bin/reth/src/lib.rs

@@ -14,6 +14,7 @@ pub mod p2p;
 pub mod prometheus_exporter;
 pub mod stage;
 pub mod test_eth_chain;
+pub mod test_vectors;
 pub use reth_staged_sync::utils;
 
 use clap::Args;

bin/reth/src/test_vectors/mod.rs

@@ -0,0 +1,33 @@
+//! Command for generating test vectors.
+use clap::{Parser, Subcommand};
+
+mod tables;
+
+/// Generate test-vectors for different data types.
+#[derive(Debug, Parser)]
+pub struct Command {
+    #[clap(subcommand)]
+    command: Subcommands,
+}
+
+#[derive(Subcommand, Debug)]
+/// `reth test-vectors` subcommands
+pub enum Subcommands {
+    /// Generates test vectors for specified tables. If no table is specified, generate for all.
+    Tables {
+        /// List of table names. Case-sensitive.
+        names: Vec<String>,
+    },
+}
+
+impl Command {
+    /// Execute the command
+    pub async fn execute(self) -> eyre::Result<()> {
+        match self.command {
+            Subcommands::Tables { names } => {
+                tables::generate_vectors(names)?;
+            }
+        }
+        Ok(())
+    }
+}

bin/reth/src/test_vectors/tables.rs

@@ -0,0 +1,170 @@
+use std::collections::HashSet;
+
+use eyre::Result;
+use proptest::{
+    arbitrary::Arbitrary,
+    prelude::{any_with, ProptestConfig},
+    strategy::{Strategy, ValueTree},
+    test_runner::TestRunner,
+};
+use reth_db::{
+    table::{DupSort, Table},
+    tables,
+};
+use tracing::error;
+
+const VECTORS_FOLDER: &str = "testdata/micro/db";
+const PER_TABLE: usize = 1000;
+
+/// Generates test vectors for specified `tables`. If list is empty, then generate for all tables.
+pub(crate) fn generate_vectors(mut tables: Vec<String>) -> Result<()> {
+    let mut runner = TestRunner::new(ProptestConfig::default());
+    std::fs::create_dir_all(VECTORS_FOLDER)?;
+
+    macro_rules! generate_vector {
+        ($table_type:ident, $per_table:expr, TABLE) => {
+            generate_table_vector::<tables::$table_type>(&mut runner, $per_table)?;
+        };
+        ($table_type:ident, $per_table:expr, DUPSORT) => {
+            generate_dupsort_vector::<tables::$table_type>(&mut runner, $per_table)?;
+        };
+    }
+
+    macro_rules! generate {
+        ([$(($table_type:ident, $per_table:expr, $table_or_dup:tt)),*]) => {
+            let all_tables = vec![$(stringify!($table_type).to_string(),)*];
+
+            if tables.is_empty() {
+                tables = all_tables;
+            }
+
+            for table in tables {
+                match table.as_str() {
+                    $(
+                        stringify!($table_type) => {
+                            println!("Generating test vectors for {} <{}>.", stringify!($table_or_dup), tables::$table_type::NAME);
+
+                            generate_vector!($table_type, $per_table, $table_or_dup);
+                        },
+                    )*
+                    _ => {
+                        error!(target: "reth::cli", "Unknown table: {}", table);
+                    }
+                }
+            }
+        }
+    }
+
+    generate!([
+        (CanonicalHeaders, PER_TABLE, TABLE),
+        (HeaderTD, PER_TABLE, TABLE),
+        (HeaderNumbers, PER_TABLE, TABLE),
+        (Headers, PER_TABLE, TABLE),
+        (BlockBodies, PER_TABLE, TABLE),
+        (BlockOmmers, 100, TABLE),
+        (TxHashNumber, PER_TABLE, TABLE),
+        (BlockTransitionIndex, PER_TABLE, TABLE),
+        (TxTransitionIndex, PER_TABLE, TABLE),
+        (Transactions, 100, TABLE),
+        (PlainStorageState, PER_TABLE, DUPSORT),
+        (PlainAccountState, PER_TABLE, TABLE)
+    ]);
+
+    Ok(())
+}
+
+/// Generates test-vectors for normal tables. Keys are sorted and not repeated.
+fn generate_table_vector<T: Table>(runner: &mut TestRunner, per_table: usize) -> Result<()>
+where
+    T::Key: Arbitrary + serde::Serialize + Ord + std::hash::Hash,
+    T::Value: Arbitrary + serde::Serialize,
+{
+    let mut rows = vec![];
+    let mut seen_keys = HashSet::new();
+    let strat = proptest::collection::vec(
+        any_with::<(T::Key, T::Value)>((
+            <T::Key as Arbitrary>::Parameters::default(),
+            <T::Value as Arbitrary>::Parameters::default(),
+        )),
+        per_table - rows.len(),
+    )
+    .no_shrink()
+    .boxed();
+
+    while rows.len() < per_table {
+        // Generate all `per_table` rows: (Key, Value)
+        rows.extend(
+            &mut strat
+                .new_tree(runner)
+                .map_err(|e| eyre::eyre!("{e}"))?
+                .current()
+                .into_iter()
+                .filter(|e| seen_keys.insert(e.0.clone())),
+        );
+    }
+    // Sort them by `Key`
+    rows.sort_by(|a, b| a.0.cmp(&b.0));
+
+    save_to_file::<T>(rows)
+}
+
+/// Generates test-vectors for DUPSORT tables. Each key has multiple (subkey, value). Keys and
+/// subkeys are sorted.
+fn generate_dupsort_vector<T: Table>(runner: &mut TestRunner, per_table: usize) -> Result<()>
+where
+    T: DupSort,
+    T::Key: Arbitrary + serde::Serialize + Ord + std::hash::Hash,
+    T::Value: Arbitrary + serde::Serialize + Ord,
+{
+    let mut rows = vec![];
+
+    // We want to control our repeated keys
+    let mut seen_keys = HashSet::new();
+
+    let strat_values = proptest::collection::vec(
+        any_with::<T::Value>(<T::Value as Arbitrary>::Parameters::default()),
+        100..300,
+    )
+    .no_shrink()
+    .boxed();
+
+    let strat_keys =
+        any_with::<T::Key>(<T::Key as Arbitrary>::Parameters::default()).no_shrink().boxed();
+
+    while rows.len() < per_table {
+        let key: T::Key = strat_keys.new_tree(runner).map_err(|e| eyre::eyre!("{e}"))?.current();
+
+        if !seen_keys.insert(key.clone()) {
+            continue
+        }
+
+        let mut values: Vec<T::Value> =
+            strat_values.new_tree(runner).map_err(|e| eyre::eyre!("{e}"))?.current();
+
+        values.sort();
+
+        for value in values {
+            rows.push((key.clone(), value));
+        }
+    }
+
+    // Sort them by `Key`
+    rows.sort_by(|a, b| a.0.cmp(&b.0));
+
+    save_to_file::<T>(rows)
+}
+
+/// Save rows to file.
+fn save_to_file<T: Table>(rows: Vec<(T::Key, T::Value)>) -> eyre::Result<()>
+where
+    T::Key: serde::Serialize,
+    T::Value: serde::Serialize,
+{
+    serde_json::to_writer_pretty(
+        std::io::BufWriter::new(
+            std::fs::File::create(format!("{VECTORS_FOLDER}/{}.json", T::NAME)).unwrap(),
+        ),
+        &rows,
+    )
+    .map_err(|e| eyre::eyre!({ e }))
+}

crates/primitives/src/storage.rs

@@ -1,9 +1,10 @@
 use super::{H256, U256};
-use reth_codecs::Compact;
-use serde::Serialize;
+use reth_codecs::{derive_arbitrary, Compact};
+use serde::{Deserialize, Serialize};
 
 /// Account storage entry.
-#[derive(Debug, Default, Copy, Clone, PartialEq, Eq, Serialize)]
+#[derive_arbitrary(compact)]
+#[derive(Debug, Default, Copy, Clone, PartialEq, Eq, Serialize, Deserialize, PartialOrd, Ord)]
 pub struct StorageEntry {
     /// Storage key.
     pub key: H256,

crates/storage/db/Cargo.toml

@@ -63,10 +63,14 @@ arbitrary = { version = "1.1.7", features = ["derive"] }
 proptest = { version = "1.0" }
 proptest-derive = "0.3"
 
+serde_json = "1.0"
+
+paste = "1.0"
+
 
 [features]
 default = ["mdbx"]
-test-utils = ["tempfile"]
+test-utils = ["tempfile", "arbitrary"]
 bench-postcard = ["bench"]
 mdbx = ["reth-libmdbx"]
 bench = []
@@ -79,9 +83,9 @@ arbitrary = [
 ]
 
 [[bench]]
-name = "encoding_crit"
+name = "criterion"
 harness = false
 
 [[bench]]
-name = "encoding_iai"
-harness = false
+name = "iai"
+harness = false

crates/storage/db/benches/criterion.rs

@@ -0,0 +1,280 @@
+#![allow(dead_code, unused_imports, non_snake_case)]
+
+use criterion::{
+    black_box, criterion_group, criterion_main, measurement::WallTime, BenchmarkGroup, Criterion,
+};
+use reth_db::cursor::{DbDupCursorRO, DbDupCursorRW};
+use std::time::Instant;
+
+criterion_group!(benches, db, serialization);
+criterion_main!(benches);
+
+pub fn db(c: &mut Criterion) {
+    let mut group = c.benchmark_group("tables_db");
+    group.measurement_time(std::time::Duration::from_millis(200));
+    group.warm_up_time(std::time::Duration::from_millis(200));
+
+    measure_table_db::<CanonicalHeaders>(&mut group);
+    measure_table_db::<HeaderTD>(&mut group);
+    measure_table_db::<HeaderNumbers>(&mut group);
+    measure_table_db::<Headers>(&mut group);
+    measure_table_db::<BlockBodies>(&mut group);
+    measure_table_db::<BlockOmmers>(&mut group);
+    measure_table_db::<TxHashNumber>(&mut group);
+    measure_table_db::<BlockTransitionIndex>(&mut group);
+    measure_table_db::<TxTransitionIndex>(&mut group);
+    measure_table_db::<Transactions>(&mut group);
+    measure_dupsort_db::<PlainStorageState>(&mut group);
+    measure_table_db::<PlainAccountState>(&mut group);
+}
+
+pub fn serialization(c: &mut Criterion) {
+    let mut group = c.benchmark_group("tables_serialization");
+    group.measurement_time(std::time::Duration::from_millis(200));
+    group.warm_up_time(std::time::Duration::from_millis(200));
+
+    measure_table_serialization::<CanonicalHeaders>(&mut group);
+    measure_table_serialization::<HeaderTD>(&mut group);
+    measure_table_serialization::<HeaderNumbers>(&mut group);
+    measure_table_serialization::<Headers>(&mut group);
+    measure_table_serialization::<BlockBodies>(&mut group);
+    measure_table_serialization::<BlockOmmers>(&mut group);
+    measure_table_serialization::<TxHashNumber>(&mut group);
+    measure_table_serialization::<BlockTransitionIndex>(&mut group);
+    measure_table_serialization::<TxTransitionIndex>(&mut group);
+    measure_table_serialization::<Transactions>(&mut group);
+    measure_table_serialization::<PlainStorageState>(&mut group);
+    measure_table_serialization::<PlainAccountState>(&mut group);
+}
+
+/// Measures `Encode`, `Decode`, `Compress` and `Decompress`.
+fn measure_table_serialization<T>(group: &mut BenchmarkGroup<WallTime>)
+where
+    T: Table + Default,
+    T::Key: Default + Clone + for<'de> serde::Deserialize<'de>,
+    T::Value: Default + Clone + for<'de> serde::Deserialize<'de>,
+{
+    let input = &load_vectors::<T>();
+    group.bench_function(format!("{}.KeyEncode", T::NAME), move |b| {
+        b.iter_with_setup(
+            || input.clone(),
+            |input| {
+                black_box({
+                    for (k, _, _, _) in input {
+                        k.encode();
+                    }
+                });
+            },
+        )
+    });
+
+    group.bench_function(format!("{}.KeyDecode", T::NAME), |b| {
+        b.iter_with_setup(
+            || input.clone(),
+            |input| {
+                black_box({
+                    for (_, k, _, _) in input {
+                        let _ = <T as Table>::Key::decode(k);
+                    }
+                });
+            },
+        )
+    });
+
+    group.bench_function(format!("{}.ValueCompress", T::NAME), move |b| {
+        b.iter_with_setup(
+            || input.clone(),
+            |input| {
+                black_box({
+                    for (_, _, v, _) in input {
+                        v.compress();
+                    }
+                });
+            },
+        )
+    });
+
+    group.bench_function(format!("{}.ValueDecompress", T::NAME), |b| {
+        b.iter_with_setup(
+            || input.clone(),
+            |input| {
+                black_box({
+                    for (_, _, _, v) in input {
+                        let _ = <T as Table>::Value::decompress(v);
+                    }
+                });
+            },
+        )
+    });
+}
+
+/// Measures `SeqWrite`, `RandomWrite`, `SeqRead` and `RandomRead` using `cursor` and `tx.put`.
+fn measure_table_db<T>(group: &mut BenchmarkGroup<WallTime>)
+where
+    T: Table + Default,
+    T::Key: Default + Clone + for<'de> serde::Deserialize<'de>,
+    T::Value: Default + Clone + for<'de> serde::Deserialize<'de>,
+{
+    let input = &load_vectors::<T>();
+    let bench_db_path = Path::new(BENCH_DB_PATH);
+
+    group.bench_function(format!("{}.SeqWrite", T::NAME), |b| {
+        b.iter_with_setup(
+            || {
+                // Reset DB
+                let _ = std::fs::remove_dir_all(bench_db_path);
+                (input.clone(), create_test_db_with_path::<WriteMap>(EnvKind::RW, bench_db_path))
+            },
+            |(input, db)| {
+                // Create TX
+                let tx = db.tx_mut().expect("tx");
+                let mut crsr = tx.cursor_write::<T>().expect("cursor");
+
+                black_box({
+                    for (k, _, v, _) in input {
+                        crsr.append(k, v).expect("submit");
+                    }
+
+                    tx.inner.commit().unwrap();
+                });
+            },
+        )
+    });
+
+    group.bench_function(format!("{}.RandomWrite", T::NAME), |b| {
+        b.iter_with_setup(
+            || {
+                // Reset DB
+                let _ = std::fs::remove_dir_all(bench_db_path);
+                (input, create_test_db_with_path::<WriteMap>(EnvKind::RW, bench_db_path))
+            },
+            |(input, db)| {
+                // Create TX
+                let tx = db.tx_mut().expect("tx");
+                let mut crsr = tx.cursor_write::<T>().expect("cursor");
+
+                black_box({
+                    for index in RANDOM_INDEXES {
+                        let (k, _, v, _) = input.get(index).unwrap().clone();
+                        crsr.insert(k, v).expect("submit");
+                    }
+
+                    tx.inner.commit().unwrap();
+                });
+            },
+        )
+    });
+
+    group.bench_function(format!("{}.SeqRead", T::NAME), |b| {
+        let db = set_up_db::<T>(bench_db_path, input);
+
+        b.iter(|| {
+            // Create TX
+            let tx = db.tx().expect("tx");
+
+            black_box({
+                let mut cursor = tx.cursor_read::<T>().expect("cursor");
+                let walker = cursor.walk(input.first().unwrap().0.clone()).unwrap();
+                for element in walker {
+                    element.unwrap();
+                }
+            });
+        })
+    });
+
+    group.bench_function(format!("{}.RandomRead", T::NAME), |b| {
+        let db = set_up_db::<T>(bench_db_path, input);
+
+        b.iter(|| {
+            // Create TX
+            let tx = db.tx().expect("tx");
+
+            black_box({
+                for index in RANDOM_INDEXES {
+                    let mut cursor = tx.cursor_read::<T>().expect("cursor");
+                    cursor.seek_exact(input.get(index).unwrap().0.clone()).unwrap();
+                }
+            });
+        })
+    });
+}
+
+/// Measures `SeqWrite`, `RandomWrite` and `SeqRead`  using `cursor_dup` and `tx.put`.
+fn measure_dupsort_db<T>(group: &mut BenchmarkGroup<WallTime>)
+where
+    T: Table + Default + DupSort,
+    T::Key: Default + Clone + for<'de> serde::Deserialize<'de>,
+    T::Value: Default + Clone + for<'de> serde::Deserialize<'de>,
+    T::SubKey: Default + Clone + for<'de> serde::Deserialize<'de>,
+{
+    let input = &load_vectors::<T>();
+    let bench_db_path = Path::new(BENCH_DB_PATH);
+
+    group.bench_function(format!("{}.SeqWrite", T::NAME), |b| {
+        b.iter_with_setup(
+            || {
+                // Reset DB
+                let _ = std::fs::remove_dir_all(bench_db_path);
+                (input.clone(), create_test_db_with_path::<WriteMap>(EnvKind::RW, bench_db_path))
+            },
+            |(input, db)| {
+                // Create TX
+                let tx = db.tx_mut().expect("tx");
+                let mut crsr = tx.cursor_dup_write::<T>().expect("cursor");
+
+                black_box({
+                    for (k, _, v, _) in input {
+                        crsr.append_dup(k, v).expect("submit");
+                    }
+
+                    tx.inner.commit().unwrap();
+                });
+            },
+        )
+    });
+
+    group.bench_function(format!("{}.RandomWrite", T::NAME), |b| {
+        b.iter_with_setup(
+            || {
+                // Reset DB
+                let _ = std::fs::remove_dir_all(bench_db_path);
+
+                (input, create_test_db_with_path::<WriteMap>(EnvKind::RW, bench_db_path))
+            },
+            |(input, db)| {
+                // Create TX
+                let tx = db.tx_mut().expect("tx");
+
+                for index in RANDOM_INDEXES {
+                    let (k, _, v, _) = input.get(index).unwrap().clone();
+                    tx.put::<T>(k, v).unwrap();
+                }
+
+                tx.inner.commit().unwrap();
+            },
+        )
+    });
+
+    group.bench_function(format!("{}.SeqRead", T::NAME), |b| {
+        let db = set_up_db::<T>(bench_db_path, input);
+
+        b.iter(|| {
+            // Create TX
+            let tx = db.tx().expect("tx");
+
+            black_box({
+                let mut cursor = tx.cursor_dup_read::<T>().expect("cursor");
+                let walker = cursor
+                    .walk_dup(input.first().unwrap().0.clone(), T::SubKey::default())
+                    .unwrap();
+                for element in walker {
+                    element.unwrap();
+                }
+            });
+        })
+    });
+
+    // group.bench_function(format!("{}.RandomRead", T::NAME), |b| {});
+}
+
+include!("./utils.rs");

crates/storage/db/benches/encoding_crit.rs

@@ -1,33 +0,0 @@
-use criterion::{black_box, criterion_group, criterion_main, Criterion};
-
-/// Benchmarks the encoding and decoding of `IntegerList` using criterion.
-macro_rules! impl_criterion_encoding_benchmark {
-    ($name:tt) => {
-        pub fn criterion_benchmark(c: &mut Criterion) {
-            let mut size = 0;
-            c.bench_function(stringify!($name), |b| {
-                b.iter(|| {
-                    let encoded_size =
-                        reth_db::tables::codecs::fuzz::IntegerList::encode_and_decode(black_box(
-                            reth_primitives::IntegerList::default(),
-                        ))
-                        .0;
-
-                    if size == 0 {
-                        size = encoded_size;
-                    }
-                })
-            });
-            println!("Size (bytes): `{size}`");
-        }
-
-        criterion_group!(benches, criterion_benchmark);
-        criterion_main!(benches);
-    };
-}
-
-#[cfg(not(feature = "bench-postcard"))]
-impl_criterion_encoding_benchmark!(scale);
-
-#[cfg(feature = "bench-postcard")]
-impl_criterion_encoding_benchmark!(postcard);

crates/storage/db/benches/encoding_iai.rs

@@ -1,21 +0,0 @@
-use iai::{black_box, main};
-use reth_db::tables::codecs;
-
-/// Benchmarks the encoding and decoding of `Header` using iai.
-macro_rules! impl_iai_encoding_benchmark {
-    ($name:tt) => {
-        fn $name() {
-            codecs::fuzz::IntegerList::encode_and_decode(black_box(
-                reth_primitives::IntegerList::default(),
-            ));
-        }
-
-        main!($name);
-    };
-}
-
-#[cfg(not(feature = "bench-postcard"))]
-impl_iai_encoding_benchmark!(scale);
-
-#[cfg(feature = "bench-postcard")]
-impl_iai_encoding_benchmark!(postcard);

crates/storage/db/benches/iai.rs

@@ -0,0 +1,79 @@
+#![allow(dead_code, unused_imports, non_snake_case)]
+
+use iai::main;
+use paste::paste;
+
+macro_rules! impl_iai_inner {
+    (
+        $(($name:tt, $mod:tt, $compress:tt, $decompress:tt, $encode:tt, $decode:tt, $seqread:tt, $randread:tt, $seqwrite:tt, $randwrite:tt))+
+    ) => {
+            $(
+                mod $mod {
+                    use iai::{black_box};
+                    include!("./utils.rs");
+
+                    pub fn $compress() {
+                        for (_, _, v, _) in black_box(load_vectors::<reth_db::tables::$name>()) {
+                            black_box(v.compress());
+                        }
+                    }
+                    pub fn $decompress() {
+                        for (_, _, _, comp) in black_box(load_vectors::<reth_db::tables::$name>()) {
+                            let _ = black_box(<reth_db::tables::$name as Table>::Value::decompress(comp));
+                        }
+                    }
+                    pub fn $encode() {
+                        for (k, _, _, _) in black_box(load_vectors::<reth_db::tables::$name>()) {
+                            black_box(k.encode());
+                        }
+                    }
+                    pub fn $decode() {
+                        for (_, enc, _, _) in black_box(load_vectors::<reth_db::tables::$name>()) {
+                            let _ = black_box(<reth_db::tables::$name as Table>::Key::decode(enc));
+                        }
+                    }
+                    pub fn $seqread() {}
+                    pub fn $randread() {}
+                    pub fn $seqwrite() {}
+                    pub fn $randwrite() {}
+                }
+                use $mod::*;
+            )+
+
+        main!(
+            $(
+                $compress,
+                $decompress,
+                $encode,
+                $decode,
+            )+
+        );
+    };
+}
+
+macro_rules! impl_iai {
+    ($($name:tt),+) => {
+            paste! {
+                impl_iai_inner!(
+                    $(
+                        ( $name, [<$name _mod>], [<$name _ValueCompress>], [<$name _ValueDecompress>], [<$name _ValueEncode>], [<$name _ValueDecode>], [<$name _SeqRead>], [<$name _RandomRead>], [<$name _SeqWrite>], [<$name _RandomWrite>])
+                    )+
+                );
+            }
+    };
+}
+
+impl_iai!(
+    CanonicalHeaders,
+    HeaderTD,
+    HeaderNumbers,
+    Headers,
+    BlockBodies,
+    BlockOmmers,
+    TxHashNumber,
+    BlockTransitionIndex,
+    TxTransitionIndex,
+    Transactions,
+    PlainStorageState,
+    PlainAccountState
+);

crates/storage/db/benches/utils.rs

@@ -0,0 +1,70 @@
+use reth_db::{
+    cursor::{DbCursorRO, DbCursorRW},
+    database::Database,
+    mdbx::{test_utils::create_test_db_with_path, EnvKind, WriteMap},
+    table::*,
+    tables::*,
+    transaction::{DbTx, DbTxMut},
+};
+use std::path::Path;
+
+/// Path where the DB is initialized for benchmarks.
+const BENCH_DB_PATH: &str = "/tmp/reth-benches";
+
+/// Used for RandomRead and RandomWrite benchmarks.
+const RANDOM_INDEXES: [usize; 10] = [23, 2, 42, 5, 3, 99, 54, 0, 33, 64];
+
+/// Returns bench vectors in the format: `Vec<(Key, EncodedKey, Value, CompressedValue)>`.
+fn load_vectors<T: reth_db::table::Table>() -> Vec<(T::Key, bytes::Bytes, T::Value, bytes::Bytes)>
+where
+    T: Default,
+    T::Key: Default + Clone + for<'de> serde::Deserialize<'de>,
+    T::Value: Default + Clone + for<'de> serde::Deserialize<'de>,
+{
+    let list: Vec<(T::Key, T::Value)> = serde_json::from_reader(std::io::BufReader::new(
+        std::fs::File::open(format!(
+            "{}/../../../testdata/micro/db/{}.json",
+            env!("CARGO_MANIFEST_DIR"),
+            T::NAME
+        ))
+        .expect("Test vectors not found. They can be generated from the workspace by calling `cargo run --bin reth -- test-vectors tables`."),
+    ))
+    .unwrap();
+
+    list.into_iter()
+        .map(|(k, v)| {
+            (
+                k.clone(),
+                bytes::Bytes::copy_from_slice(k.encode().as_ref()),
+                v.clone(),
+                bytes::Bytes::copy_from_slice(v.compress().as_ref()),
+            )
+        })
+        .collect::<Vec<_>>()
+}
+
+/// Sets up a clear database at `bench_db_path`.
+fn set_up_db<T>(
+    bench_db_path: &Path,
+    pair: &Vec<(<T as Table>::Key, bytes::Bytes, <T as Table>::Value, bytes::Bytes)>,
+) -> reth_db::mdbx::Env<WriteMap>
+where
+    T: Table + Default,
+    T::Key: Default + Clone,
+    T::Value: Default + Clone,
+{
+    // Reset DB
+    let _ = std::fs::remove_dir_all(bench_db_path);
+    let db = create_test_db_with_path::<WriteMap>(EnvKind::RW, bench_db_path);
+
+    {
+        // Prepare data to be read
+        let tx = db.tx_mut().expect("tx");
+        for (k, _, v, _) in pair.clone() {
+            tx.put::<T>(k, v).expect("submit");
+        }
+        tx.inner.commit().unwrap();
+    }
+
+    db
+}

crates/storage/db/src/tables/models/blocks.rs

@@ -65,7 +65,7 @@ pub type HeaderHash = H256;
 /// element as BlockNumber, helps out with querying/sorting.
 ///
 /// Since it's used as a key, the `BlockNumber` is not compressed when encoding it.
-#[derive(Clone, Copy, PartialEq, Eq, Serialize, Deserialize, Default, Ord, PartialOrd)]
+#[derive(Clone, Copy, PartialEq, Eq, Serialize, Deserialize, Default, Ord, PartialOrd, Hash)]
 pub struct BlockNumHash(pub (BlockNumber, BlockHash));
 
 impl std::fmt::Debug for BlockNumHash {

crates/storage/db/src/tables/utils.rs

@@ -23,6 +23,20 @@ macro_rules! impl_fixed_arbitrary {
                 Decode::decode(buffer).map_err(|_| arbitrary::Error::IncorrectFormat)
             }
         }
+
+        #[cfg(any(test, feature = "arbitrary"))]
+        use proptest::strategy::Strategy;
+        #[cfg(any(test, feature = "arbitrary"))]
+        impl proptest::prelude::Arbitrary for $name {
+            type Parameters = ();
+            type Strategy = proptest::prelude::BoxedStrategy<$name>;
+
+            fn arbitrary_with(args: Self::Parameters) -> Self::Strategy {
+                proptest::collection::vec(proptest::arbitrary::any_with::<u8>(args), $size)
+                    .prop_map(move |vec| Decode::decode(vec).unwrap())
+                    .boxed()
+            }
+        }
     };
 }