blob: ce4f5e507a2ce4c1b4edc1a66e6a6b9f300c5906 [file] [log] [blame]
/*
*
* Copyright 2019 gRPC authors.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package wrr
import (
"errors"
"math"
"math/rand"
"strconv"
"testing"
"github.com/google/go-cmp/cmp"
"google.golang.org/grpc/internal/grpctest"
)
type s struct {
grpctest.Tester
}
func Test(t *testing.T) {
grpctest.RunSubTests(t, s{})
}
const iterCount = 10000
func equalApproximate(a, b float64) error {
opt := cmp.Comparer(func(x, y float64) bool {
delta := math.Abs(x - y)
mean := math.Abs(x+y) / 2.0
return delta/mean < 0.05
})
if !cmp.Equal(a, b, opt) {
return errors.New(cmp.Diff(a, b))
}
return nil
}
func testWRRNext(t *testing.T, newWRR func() WRR) {
tests := []struct {
name string
weights []int64
}{
{
name: "1-1-1",
weights: []int64{1, 1, 1},
},
{
name: "1-2-3",
weights: []int64{1, 2, 3},
},
{
name: "5-3-2",
weights: []int64{5, 3, 2},
},
{
name: "17-23-37",
weights: []int64{17, 23, 37},
},
{
name: "no items",
weights: []int64{},
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
w := newWRR()
if len(tt.weights) == 0 {
if next := w.Next(); next != nil {
t.Fatalf("w.Next returns non nil value:%v when there is no item", next)
}
return
}
var sumOfWeights int64
for i, weight := range tt.weights {
w.Add(i, weight)
sumOfWeights += weight
}
results := make(map[int]int)
for i := 0; i < iterCount; i++ {
results[w.Next().(int)]++
}
wantRatio := make([]float64, len(tt.weights))
for i, weight := range tt.weights {
wantRatio[i] = float64(weight) / float64(sumOfWeights)
}
gotRatio := make([]float64, len(tt.weights))
for i, count := range results {
gotRatio[i] = float64(count) / iterCount
}
for i := range wantRatio {
if err := equalApproximate(gotRatio[i], wantRatio[i]); err != nil {
t.Errorf("%v not equal %v", i, err)
}
}
})
}
}
func (s) TestRandomWRRNext(t *testing.T) {
testWRRNext(t, NewRandom)
}
func (s) TestEdfWrrNext(t *testing.T) {
testWRRNext(t, NewEDF)
}
func BenchmarkRandomWRRNext(b *testing.B) {
for _, n := range []int{100, 500, 1000} {
b.Run("equal-weights-"+strconv.Itoa(n)+"-items", func(b *testing.B) {
w := NewRandom()
sumOfWeights := n
for i := 0; i < n; i++ {
w.Add(i, 1)
}
b.ResetTimer()
for i := 0; i < b.N; i++ {
for i := 0; i < sumOfWeights; i++ {
w.Next()
}
}
})
}
var maxWeight int64 = 1024
for _, n := range []int{100, 500, 1000} {
b.Run("random-weights-"+strconv.Itoa(n)+"-items", func(b *testing.B) {
w := NewRandom()
var sumOfWeights int64
for i := 0; i < n; i++ {
weight := rand.Int63n(maxWeight + 1)
w.Add(i, weight)
sumOfWeights += weight
}
b.ResetTimer()
for i := 0; i < b.N; i++ {
for i := 0; i < int(sumOfWeights); i++ {
w.Next()
}
}
})
}
itemsNum := 200
heavyWeight := int64(itemsNum)
lightWeight := int64(1)
heavyIndices := []int{0, itemsNum / 2, itemsNum - 1}
for _, heavyIndex := range heavyIndices {
b.Run("skew-weights-heavy-index-"+strconv.Itoa(heavyIndex), func(b *testing.B) {
w := NewRandom()
var sumOfWeights int64
for i := 0; i < itemsNum; i++ {
var weight int64
if i == heavyIndex {
weight = heavyWeight
} else {
weight = lightWeight
}
sumOfWeights += weight
w.Add(i, weight)
}
b.ResetTimer()
for i := 0; i < b.N; i++ {
for i := 0; i < int(sumOfWeights); i++ {
w.Next()
}
}
})
}
}
func init() {
r := rand.New(rand.NewSource(0))
grpcrandInt63n = r.Int63n
}