Add parity-db variant for trie benchmarks (#5827)

* parity-db bench

* use arkady suggestion

substrate/bin/node/bench/Cargo.toml

@@ -30,3 +30,5 @@ fs_extra = "1"
 hex = "0.4.0"
 rand = { version = "0.7.2", features = ["small_rng"] }
 lazy_static = "1.4.0"
+parity-util-mem = { version = "0.6.1", default-features = false, features = ["primitive-types"] }
+parity-db = { version = "0.1.2" }

substrate/bin/node/bench/src/core.rs

@@ -139,10 +139,10 @@ pub fn run_benchmark(
 }
 
 macro_rules! matrix(
-	( $var:ident in $over:expr => $tt:expr,  $( $rest:tt )* ) => {
+	( $var:tt in $over:expr => $tt:expr,  $( $rest:tt )* ) => {
 		{
 			let mut res = Vec::<Box<dyn crate::core::BenchmarkDescription>>::new();
-			for $var in $over.iter() {
+			for $var in $over {
 				res.push(Box::new($tt));
 			}
 			res.extend(matrix!( $($rest)* ));

substrate/bin/node/bench/src/main.rs

@@ -23,6 +23,7 @@ mod tempdb;
 mod state_sizes;
 
 use crate::core::{run_benchmark, Mode as BenchmarkMode};
+use crate::tempdb::DatabaseType;
 use import::{ImportBenchmarkDescription, SizeType};
 use trie::{TrieReadBenchmarkDescription, TrieWriteBenchmarkDescription, DatabaseSize};
 use node_testing::bench::{Profile, KeyTypes};
@@ -66,7 +67,7 @@ fn main() {
 	}
 
 	let benchmarks = matrix!(
-		profile in [Profile::Wasm, Profile::Native] =>
+		profile in [Profile::Wasm, Profile::Native].iter() =>
 			ImportBenchmarkDescription {
 				profile: *profile,
 				key_types: KeyTypes::Sr25519,
@@ -87,22 +88,36 @@ fn main() {
 			key_types: KeyTypes::Sr25519,
 			size: SizeType::Full,
 		},
-		size in [SizeType::Small, SizeType::Large] =>
+		size in [SizeType::Small, SizeType::Large].iter() =>
 			ImportBenchmarkDescription {
 				profile: Profile::Native,
 				key_types: KeyTypes::Sr25519,
 				size: *size,
 			},
-		size in [
-			DatabaseSize::Empty, DatabaseSize::Smallest, DatabaseSize::Small,
-			DatabaseSize::Medium, DatabaseSize::Large, DatabaseSize::Huge,
-		] => TrieReadBenchmarkDescription { database_size: *size },
-		size in [
-			DatabaseSize::Empty, DatabaseSize::Smallest, DatabaseSize::Small,
-			DatabaseSize::Medium, DatabaseSize::Large, DatabaseSize::Huge,
-		] => TrieWriteBenchmarkDescription { database_size: *size },
+		(size, db_type) in
+			[
+				DatabaseSize::Empty, DatabaseSize::Smallest, DatabaseSize::Small,
+				DatabaseSize::Medium, DatabaseSize::Large, DatabaseSize::Huge,
+			]
+			.iter().flat_map(|size|
+			[
+				DatabaseType::RocksDb, DatabaseType::ParityDb
+			]
+			.iter().map(move |db_type| (size, db_type)))
+			=> TrieReadBenchmarkDescription { database_size: *size, database_type: *db_type },
+		(size, db_type) in
+			[
+				DatabaseSize::Empty, DatabaseSize::Smallest, DatabaseSize::Small,
+				DatabaseSize::Medium, DatabaseSize::Large, DatabaseSize::Huge,
+			]
+			.iter().flat_map(|size|
+			[
+				DatabaseType::RocksDb, DatabaseType::ParityDb
+			]
+			.iter().map(move |db_type| (size, db_type)))
+			=> TrieWriteBenchmarkDescription { database_size: *size, database_type: *db_type },
 	);
-	
+
 	if opt.list {
 		for benchmark in benchmarks.iter() {
 			log::info!("{}: {}", benchmark.name(), benchmark.path().full())

substrate/bin/node/bench/src/tempdb.rs

@@ -14,12 +14,67 @@
 // You should have received a copy of the GNU General Public License
 // along with Substrate.  If not, see <http://www.gnu.org/licenses/>.
 
-use std::sync::Arc;
-use kvdb::KeyValueDB;
+use std::{io, sync::Arc};
+use kvdb::{KeyValueDB, DBTransaction};
 use kvdb_rocksdb::{DatabaseConfig, Database};
 
+#[derive(Debug, Clone, Copy, derive_more::Display)]
+pub enum DatabaseType {
+	RocksDb,
+	ParityDb,
+}
+
 pub struct TempDatabase(tempfile::TempDir);
 
+struct ParityDbWrapper(parity_db::Db);
+parity_util_mem::malloc_size_of_is_0!(ParityDbWrapper);
+
+impl KeyValueDB for ParityDbWrapper {
+	/// Get a value by key.
+	fn get(&self, col: u32, key: &[u8]) -> io::Result<Option<Vec<u8>>> {
+		Ok(self.0.get(col as u8, &key[key.len() - 32..]).expect("db error"))
+	}
+
+	/// Get a value by partial key. Only works for flushed data.
+	fn get_by_prefix(&self, _col: u32, _prefix: &[u8]) -> Option<Box<[u8]>> {
+		unimplemented!()
+	}
+
+	/// Write a transaction of changes to the buffer.
+	fn write_buffered(&self, transaction: DBTransaction) {
+		self.0.commit(
+			transaction.ops.iter().map(|op| match op {
+				kvdb::DBOp::Insert { col, key, value } => (*col as u8, &key[key.len() - 32..], Some(value.to_vec())),
+				kvdb::DBOp::Delete { col, key } => (*col as u8, &key[key.len() - 32..], None),
+			})
+		).expect("db error");
+	}
+
+	/// Flush all buffered data.
+	fn flush(&self) -> io::Result<()> {
+		Ok(())
+	}
+
+	/// Iterate over flushed data for a given column.
+	fn iter<'a>(&'a self, _col: u32) -> Box<dyn Iterator<Item = (Box<[u8]>, Box<[u8]>)> + 'a> {
+		unimplemented!()
+	}
+
+	/// Iterate over flushed data for a given column, starting from a given prefix.
+	fn iter_from_prefix<'a>(
+		&'a self,
+		_col: u32,
+		_prefix: &'a [u8],
+	) -> Box<dyn Iterator<Item = (Box<[u8]>, Box<[u8]>)> + 'a> {
+		unimplemented!()
+	}
+
+	/// Attempt to replace this database with a new one located at the given path.
+	fn restore(&self, _new_db: &str) -> io::Result<()> {
+		unimplemented!()
+	}
+}
+
 impl TempDatabase {
 	pub fn new() -> Self {
 		let dir = tempfile::tempdir().expect("temp dir creation failed");
@@ -32,10 +87,25 @@ impl TempDatabase {
 		TempDatabase(dir)
 	}
 
-	pub fn open(&mut self) -> Arc<dyn KeyValueDB> {
-		let db_cfg = DatabaseConfig::with_columns(1);
-		let db = Database::open(&db_cfg, &self.0.path().to_string_lossy()).expect("Database backend error");
-		Arc::new(db)
+	pub fn open(&mut self, db_type: DatabaseType) -> Arc<dyn KeyValueDB> {
+		match db_type {
+			DatabaseType::RocksDb => {
+				let db_cfg = DatabaseConfig::with_columns(1);
+				let db = Database::open(&db_cfg, &self.0.path().to_string_lossy()).expect("Database backend error");
+				Arc::new(db)
+			},
+			DatabaseType::ParityDb => {
+				Arc::new(ParityDbWrapper({
+					let mut options = parity_db::Options::with_columns(self.0.path(), 1);
+					let mut column_options = &mut options.columns[0];
+					column_options.ref_counted = true;
+					column_options.preimage = true;
+					column_options.uniform = true;
+					parity_db::Db::open(&options).expect("db open error")
+				}))
+			}
+		}
+
 	}
 }
 

substrate/bin/node/bench/src/trie.rs

@@ -30,7 +30,7 @@ use crate::{
 	core::{self, Mode, Path},
 	generator::generate_trie,
 	simple_trie::SimpleTrie,
-	tempdb::TempDatabase,
+	tempdb::{TempDatabase, DatabaseType},
 };
 
 pub const SAMPLE_SIZE: usize = 100;
@@ -91,6 +91,7 @@ fn pretty_print(v: usize) -> String {
 
 pub struct TrieReadBenchmarkDescription {
 	pub database_size: DatabaseSize,
+	pub database_type: DatabaseType,
 }
 
 pub struct TrieReadBenchmark {
@@ -98,6 +99,7 @@ pub struct TrieReadBenchmark {
 	root: Hash,
 	warmup_keys: KeyValues,
 	query_keys: KeyValues,
+	database_type: DatabaseType,
 }
 
 impl core::BenchmarkDescription for TrieReadBenchmarkDescription {
@@ -139,7 +141,7 @@ impl core::BenchmarkDescription for TrieReadBenchmarkDescription {
 		assert_eq!(query_keys.len(), SAMPLE_SIZE);
 
 		let root = generate_trie(
-			database.open(),
+			database.open(self.database_type),
 			key_values,
 		);
 
@@ -148,14 +150,16 @@ impl core::BenchmarkDescription for TrieReadBenchmarkDescription {
 			root,
 			warmup_keys,
 			query_keys,
+			database_type: self.database_type,
 		})
 	}
 
 	fn name(&self) -> Cow<'static, str> {
 		format!(
-			"Trie read benchmark({} database ({} keys))",
+			"Trie read benchmark({} database ({} keys), db_type: {})",
 			self.database_size,
 			pretty_print(self.database_size.keys()),
+			self.database_type,
 		).into()
 	}
 }
@@ -173,7 +177,7 @@ impl core::Benchmark for TrieReadBenchmark {
 	fn run(&mut self, mode: Mode) -> std::time::Duration {
 		let mut db = self.database.clone();
 		let storage: Arc<dyn sp_state_machine::Storage<sp_core::Blake2Hasher>> =
-			Arc::new(Storage(db.open()));
+			Arc::new(Storage(db.open(self.database_type)));
 
 		let trie_backend = sp_state_machine::TrieBackend::new(
 			storage,
@@ -208,8 +212,10 @@ impl core::Benchmark for TrieReadBenchmark {
 
 pub struct TrieWriteBenchmarkDescription {
 	pub database_size: DatabaseSize,
+	pub database_type: DatabaseType,
 }
 
+
 impl core::BenchmarkDescription for TrieWriteBenchmarkDescription {
 	fn path(&self) -> Path {
 		let mut path = Path::new(&["trie", "write"]);
@@ -245,7 +251,7 @@ impl core::BenchmarkDescription for TrieWriteBenchmarkDescription {
 		assert_eq!(warmup_keys.len(), SAMPLE_SIZE);
 
 		let root = generate_trie(
-			database.open(),
+			database.open(self.database_type),
 			key_values,
 		);
 
@@ -253,14 +259,16 @@ impl core::BenchmarkDescription for TrieWriteBenchmarkDescription {
 			database,
 			root,
 			warmup_keys,
+			database_type: self.database_type,
 		})
 	}
 
 	fn name(&self) -> Cow<'static, str> {
 		format!(
-			"Trie write benchmark({} database ({} keys))",
+			"Trie write benchmark({} database ({} keys), db_type = {})",
 			self.database_size,
 			pretty_print(self.database_size.keys()),
+			self.database_type,
 		).into()
 	}
 }
@@ -269,13 +277,14 @@ struct TrieWriteBenchmark {
 	database: TempDatabase,
 	root: Hash,
 	warmup_keys: KeyValues,
+	database_type: DatabaseType,
 }
 
 impl core::Benchmark for TrieWriteBenchmark {
 	fn run(&mut self, mode: Mode) -> std::time::Duration {
 		let mut rng = rand::thread_rng();
 		let mut db = self.database.clone();
-		let kvdb = db.open();
+		let kvdb = db.open(self.database_type);
 
 		let mut new_root = self.root.clone();