ethdb: add benchmark test suite (#26659)

Author: rjl493456442

ethdb/dbtest/testsuite.go

@@ -18,6 +18,7 @@ package dbtest
 
 import (
 	"bytes"
+	"math/rand"
 	"reflect"
 	"sort"
 	"testing"
@@ -377,6 +378,101 @@ func TestDatabaseSuite(t *testing.T, New func() ethdb.KeyValueStore) {
 	})
 }
 
+// BenchDatabaseSuite runs a suite of benchmarks against a KeyValueStore database
+// implementation.
+func BenchDatabaseSuite(b *testing.B, New func() ethdb.KeyValueStore) {
+	var (
+		keys, vals   = makeDataset(1_000_000, 32, 32, false)
+		sKeys, sVals = makeDataset(1_000_000, 32, 32, true)
+	)
+	// Run benchmarks sequentially
+	b.Run("Write", func(b *testing.B) {
+		benchWrite := func(b *testing.B, keys, vals [][]byte) {
+			b.ResetTimer()
+			b.ReportAllocs()
+
+			db := New()
+			defer db.Close()
+
+			for i := 0; i < len(keys); i++ {
+				db.Put(keys[i], vals[i])
+			}
+		}
+		b.Run("WriteSorted", func(b *testing.B) {
+			benchWrite(b, sKeys, sVals)
+		})
+		b.Run("WriteRandom", func(b *testing.B) {
+			benchWrite(b, keys, vals)
+		})
+	})
+	b.Run("Read", func(b *testing.B) {
+		benchRead := func(b *testing.B, keys, vals [][]byte) {
+			db := New()
+			defer db.Close()
+
+			for i := 0; i < len(keys); i++ {
+				db.Put(keys[i], vals[i])
+			}
+			b.ResetTimer()
+			b.ReportAllocs()
+
+			for i := 0; i < len(keys); i++ {
+				db.Get(keys[i])
+			}
+		}
+		b.Run("ReadSorted", func(b *testing.B) {
+			benchRead(b, sKeys, sVals)
+		})
+		b.Run("ReadRandom", func(b *testing.B) {
+			benchRead(b, keys, vals)
+		})
+	})
+	b.Run("Iteration", func(b *testing.B) {
+		benchIteration := func(b *testing.B, keys, vals [][]byte) {
+			db := New()
+			defer db.Close()
+
+			for i := 0; i < len(keys); i++ {
+				db.Put(keys[i], vals[i])
+			}
+			b.ResetTimer()
+			b.ReportAllocs()
+
+			it := db.NewIterator(nil, nil)
+			for it.Next() {
+			}
+			it.Release()
+		}
+		b.Run("IterationSorted", func(b *testing.B) {
+			benchIteration(b, sKeys, sVals)
+		})
+		b.Run("IterationRandom", func(b *testing.B) {
+			benchIteration(b, keys, vals)
+		})
+	})
+	b.Run("BatchWrite", func(b *testing.B) {
+		benchBatchWrite := func(b *testing.B, keys, vals [][]byte) {
+			b.ResetTimer()
+			b.ReportAllocs()
+
+			db := New()
+			defer db.Close()
+
+			batch := db.NewBatch()
+			for i := 0; i < len(keys); i++ {
+				batch.Put(keys[i], vals[i])
+			}
+			batch.Write()
+		}
+		b.Run("BenchWriteSorted", func(b *testing.B) {
+			benchBatchWrite(b, sKeys, sVals)
+		})
+		b.Run("BenchWriteRandom", func(b *testing.B) {
+			benchBatchWrite(b, keys, vals)
+		})
+	})
+}
+
 func iterateKeys(it ethdb.Iterator) []string {
 	keys := []string{}
 	for it.Next() {
@@ -386,3 +482,25 @@ func iterateKeys(it ethdb.Iterator) []string {
 	it.Release()
 	return keys
 }
+
+// randomHash generates a random blob of data and returns it as a hash.
+func randBytes(len int) []byte {
+	buf := make([]byte, len)
+	if n, err := rand.Read(buf); n != len || err != nil {
+		panic(err)
+	}
+	return buf
+}
+
+func makeDataset(size, ksize, vsize int, order bool) ([][]byte, [][]byte) {
+	var keys [][]byte
+	var vals [][]byte
+	for i := 0; i < size; i += 1 {
+		keys = append(keys, randBytes(ksize))
+		vals = append(vals, randBytes(vsize))
+	}
+	if order {
+		sort.Slice(keys, func(i, j int) bool { return bytes.Compare(keys[i], keys[j]) < 0 })
+	}
+	return keys, vals
+}

ethdb/leveldb/leveldb_test.go

@@ -38,3 +38,15 @@ func TestLevelDB(t *testing.T) {
 		})
 	})
 }
+
+func BenchmarkLevelDB(b *testing.B) {
+	dbtest.BenchDatabaseSuite(b, func() ethdb.KeyValueStore {
+		db, err := leveldb.Open(storage.NewMemStorage(), nil)
+		if err != nil {
+			b.Fatal(err)
+		}
+		return &Database{
+			db: db,
+		}
+	})
+}

ethdb/pebble/pebble.go

@@ -272,7 +272,9 @@ func (d *Database) NewBatch() ethdb.Batch {
 }
 
 // NewBatchWithSize creates a write-only database batch with pre-allocated buffer.
-// TODO can't do this with pebble.  Batches are allocated in a pool so maybe this doesn't matter?
+// It's not supported by pebble, but pebble has better memory allocation strategy
+// which turns out a lot faster than leveldb. It's performant enough to construct
+// batch object without any pre-allocated space.
 func (d *Database) NewBatchWithSize(_ int) ethdb.Batch {
 	return &batch{
 		b: d.db.NewBatch(),

ethdb/pebble/pebble_test.go

@@ -42,3 +42,17 @@ func TestPebbleDB(t *testing.T) {
 		})
 	})
 }
+
+func BenchmarkPebbleDB(b *testing.B) {
+	dbtest.BenchDatabaseSuite(b, func() ethdb.KeyValueStore {
+		db, err := pebble.Open("", &pebble.Options{
+			FS: vfs.NewMem(),
+		})
+		if err != nil {
+			b.Fatal(err)
+		}
+		return &Database{
+			db: db,
+		}
+	})
+}