- Simple
- No Dependencies.
- Straightforward API.
- Fast
- Outperforms well-known LRU caches.
- Zero memory allocations.
- GC friendly
- Pointerless and continuous data structs.
- Minimized GC scan times.
- Memory efficient
- Adds only 26 extra bytes per entry.
- Minimized memory usage.
- Feature optional
- Using SlidingCache via
WithSliding(true)option. - Create LoadingCache via
WithLoader(func(context.Context, K) (V, time.Duration, error))option.
- Using SlidingCache via
- The TTL is accurate to the nearest second.
- Expired items are only removed when accessed again or the cache is full.
package main
import (
"time"
"github.com/phuslu/lru"
)
func main() {
cache := lru.NewTTLCache[string, int](8192)
cache.Set("a", 1, 2*time.Second)
println(cache.Get("a"))
println(cache.Get("b"))
time.Sleep(1 * time.Second)
println(cache.Get("a"))
time.Sleep(2 * time.Second)
println(cache.Get("a"))
stats := cache.Stats()
println("SetCalls", stats.SetCalls, "GetCalls", stats.GetCalls, "Misses", stats.Misses)
}Disclaimer: This have been testing on my 3 environments and the results may be very different from yours. see #14
A Performance result as below. Check github benchmark action for more results and details.
go1.24 benchmark on keysize=16, itemsize=1000000, cachesize=50%, concurrency=8
// env writeratio=0.1 zipfian=false go test -v -cpu=8 -run=none -bench=. -benchtime=5s -benchmem bench_test.go
package bench
import (
"crypto/sha1"
"fmt"
"math/rand/v2"
"math/bits"
"os"
"runtime"
"strconv"
"testing"
"time"
_ "unsafe"
theine "github.com/Yiling-J/theine-go"
"github.com/cespare/xxhash/v2"
cloudflare "github.com/cloudflare/golibs/lrucache"
ristretto "github.com/dgraph-io/ristretto/v2"
freelru "github.com/elastic/go-freelru"
hashicorp "github.com/hashicorp/golang-lru/v2/expirable"
ccache "github.com/karlseguin/ccache/v3"
lxzan "github.com/lxzan/memorycache"
otter "github.com/maypok86/otter"
ecache "github.com/orca-zhang/ecache"
phuslu "github.com/phuslu/lru"
"github.com/aclements/go-perfevent/perfbench"
)
const (
keysize = 16
cachesize = 1000000
)
var writeratio, _ = strconv.ParseFloat(os.Getenv("writeratio"), 64)
var zipfian, _ = strconv.ParseBool(os.Getenv("zipfian"))
type CheapRand struct {
Seed uint64
}
func (rand *CheapRand) Uint32() uint32 {
rand.Seed += 0xa0761d6478bd642f
hi, lo := bits.Mul64(rand.Seed, rand.Seed^0xe7037ed1a0b428db)
return uint32(hi ^ lo)
}
func (rand *CheapRand) Uint32n(n uint32) uint32 {
return uint32((uint64(rand.Uint32()) * uint64(n)) >> 32)
}
func (rand *CheapRand) Uint64() uint64 {
return uint64(rand.Uint32())<<32 ^ uint64(rand.Uint32())
}
var shardcount = func() int {
n := runtime.GOMAXPROCS(0) * 16
k := 1
for k < n {
k = k * 2
}
return k
}()
var keys = func() (x []string) {
x = make([]string, cachesize)
for i := range cachesize {
x[i] = fmt.Sprintf("%x", sha1.Sum([]byte(fmt.Sprint(i))))[:keysize]
}
return
}()
func BenchmarkHashicorpSetGet(b *testing.B) {
c := perfbench.Open(b)
cache := hashicorp.NewLRU[string, int](cachesize, nil, time.Hour)
for i := range cachesize/2 {
cache.Add(keys[i], i)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.Add(keys[i], i)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func BenchmarkCloudflareSetGet(b *testing.B) {
c := perfbench.Open(b)
cache := cloudflare.NewMultiLRUCache(uint(shardcount), uint(cachesize/shardcount))
for i := range cachesize/2 {
cache.Set(keys[i], i, time.Now().Add(time.Hour))
}
expires := time.Now().Add(time.Hour)
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.Set(keys[i], i, expires)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func BenchmarkEcacheSetGet(b *testing.B) {
c := perfbench.Open(b)
cache := ecache.NewLRUCache(uint16(shardcount), uint16(cachesize/shardcount), time.Hour)
for i := range cachesize/2 {
cache.Put(keys[i], i)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.Put(keys[i], i)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func BenchmarkLxzanSetGet(b *testing.B) {
c := perfbench.Open(b)
cache := lxzan.New[string, int](
lxzan.WithBucketNum(shardcount),
lxzan.WithBucketSize(cachesize/shardcount, cachesize/shardcount),
lxzan.WithInterval(time.Hour, time.Hour),
)
for i := range cachesize/2 {
cache.Set(keys[i], i, time.Hour)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.Set(keys[i], i, time.Hour)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func hashStringXXHASH(s string) uint32 {
return uint32(xxhash.Sum64String(s))
}
func BenchmarkFreelruSetGet(b *testing.B) {
c := perfbench.Open(b)
cache, _ := freelru.NewSharded[string, int](cachesize, hashStringXXHASH)
for i := range cachesize/2 {
cache.AddWithLifetime(keys[i], i, time.Hour)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.AddWithLifetime(keys[i], i, time.Hour)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func BenchmarkPhusluSetGet(b *testing.B) {
c := perfbench.Open(b)
cache := phuslu.NewTTLCache[string, int](cachesize, phuslu.WithShards[string, int](uint32(shardcount)))
for i := range cachesize/2 {
cache.Set(keys[i], i, time.Hour)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.Set(keys[i], i, time.Hour)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func BenchmarkNoTTLSetGet(b *testing.B) {
c := perfbench.Open(b)
cache := phuslu.NewLRUCache[string, int](cachesize, phuslu.WithShards[string, int](uint32(shardcount)))
for i := range cachesize/2 {
cache.Set(keys[i], i)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.Set(keys[i], i)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func BenchmarkCcacheSetGet(b *testing.B) {
c := perfbench.Open(b)
cache := ccache.New(ccache.Configure[int]().MaxSize(cachesize).ItemsToPrune(100))
for i := range cachesize/2 {
cache.Set(keys[i], i, time.Hour)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.Set(keys[i], i, time.Hour)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func BenchmarkRistrettoSetGet(b *testing.B) {
c := perfbench.Open(b)
cache, _ := ristretto.NewCache(&ristretto.Config[string, int]{
NumCounters: 10 * cachesize, // number of keys to track frequency of (10M).
MaxCost: cachesize, // maximum cost of cache (1M).
BufferItems: 64, // number of keys per Get buffer.
})
for i := range cachesize/2 {
cache.SetWithTTL(keys[i], i, 1, time.Hour)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.SetWithTTL(keys[i], i, 1, time.Hour)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func BenchmarkTheineSetGet(b *testing.B) {
c := perfbench.Open(b)
cache, _ := theine.NewBuilder[string, int](cachesize).Build()
for i := range cachesize/2 {
cache.SetWithTTL(keys[i], i, 1, time.Hour)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.SetWithTTL(keys[i], i, 1, time.Hour)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}
func BenchmarkOtterSetGet(b *testing.B) {
c := perfbench.Open(b)
cache, _ := otter.MustBuilder[string, int](cachesize).WithVariableTTL().Build()
for i := range cachesize/2 {
cache.Set(keys[i], i, time.Hour)
}
b.ResetTimer()
c.Reset()
b.RunParallel(func(pb *testing.PB) {
threshold := uint32(float64(^uint32(0)) * writeratio)
cheaprand := &CheapRand{uint64(time.Now().UnixNano())}
zipf := rand.NewZipf(rand.New(cheaprand), 1.0001, 10, cachesize-1)
for pb.Next() {
if threshold > 0 && cheaprand.Uint32() <= threshold {
i := int(cheaprand.Uint32n(cachesize))
cache.Set(keys[i], i, time.Hour)
} else if zipfian {
cache.Get(keys[zipf.Uint64()])
} else {
cache.Get(keys[cheaprand.Uint32n(cachesize)])
}
}
})
}Run with 95% zipf reads and 5% randomly writes.
In github actions:
goos: linux
goarch: amd64
cpu: AMD EPYC 7763 64-Core Processor
BenchmarkHashicorpSetGet
BenchmarkHashicorpSetGet-8 14216642 437.8 ns/op 1 B/op 0 allocs/op
BenchmarkCloudflareSetGet
BenchmarkCloudflareSetGet-8 48545095 138.4 ns/op 16 B/op 1 allocs/op
BenchmarkEcacheSetGet
BenchmarkEcacheSetGet-8 58657796 111.3 ns/op 1 B/op 0 allocs/op
BenchmarkLxzanSetGet
BenchmarkLxzanSetGet-8 58624521 108.2 ns/op 0 B/op 0 allocs/op
BenchmarkFreelruSetGet
BenchmarkFreelruSetGet-8 57877179 114.8 ns/op 0 B/op 0 allocs/op
BenchmarkPhusluSetGet
BenchmarkPhusluSetGet-8 80641694 84.63 ns/op 0 B/op 0 allocs/op
BenchmarkNoTTLSetGet
BenchmarkNoTTLSetGet-8 82049758 81.10 ns/op 0 B/op 0 allocs/op
BenchmarkCcacheSetGet
BenchmarkCcacheSetGet-8 15654375 341.5 ns/op 25 B/op 2 allocs/op
BenchmarkRistrettoSetGet
BenchmarkRistrettoSetGet-8 55431110 111.9 ns/op 8 B/op 0 allocs/op
BenchmarkTheineSetGet
BenchmarkTheineSetGet-8 38325188 145.9 ns/op 2 B/op 0 allocs/op
BenchmarkOtterSetGet
BenchmarkOtterSetGet-8 70208274 93.29 ns/op 4 B/op 0 allocs/op
PASS
ok command-line-arguments 106.169s
In my windows laptop:
goos: windows
goarch: amd64
cpu: 11th Gen Intel(R) Core(TM) i5-1135G7 @ 2.40GHz
BenchmarkHashicorpSetGet
BenchmarkHashicorpSetGet-8 17376360 570.8 ns/op 0 B/op 0 allocs/op
BenchmarkCloudflareSetGet
BenchmarkCloudflareSetGet-8 74426450 92.55 ns/op 16 B/op 1 allocs/op
BenchmarkEcacheSetGet
BenchmarkEcacheSetGet-8 88482524 78.37 ns/op 1 B/op 0 allocs/op
BenchmarkLxzanSetGet
BenchmarkLxzanSetGet-8 74142448 90.51 ns/op 0 B/op 0 allocs/op
BenchmarkFreelruSetGet
BenchmarkFreelruSetGet-8 80740551 88.70 ns/op 0 B/op 0 allocs/op
BenchmarkPhusluSetGet
BenchmarkPhusluSetGet-8 97101519 65.88 ns/op 0 B/op 0 allocs/op
BenchmarkNoTTLSetGet
BenchmarkNoTTLSetGet-8 92839735 65.27 ns/op 0 B/op 0 allocs/op
BenchmarkCcacheSetGet
BenchmarkCcacheSetGet-8 25047673 350.2 ns/op 25 B/op 2 allocs/op
BenchmarkRistrettoSetGet
BenchmarkRistrettoSetGet-8 90289784 81.44 ns/op 22 B/op 1 allocs/op
BenchmarkTheineSetGet
BenchmarkTheineSetGet-8 57348163 114.3 ns/op 0 B/op 0 allocs/op
BenchmarkOtterSetGet
BenchmarkOtterSetGet-8 100000000 65.72 ns/op 4 B/op 0 allocs/op
PASS
ok command-line-arguments 104.553s
In an idle 48c256g server:
goos: linux
goarch: amd64
cpu: Intel(R) Xeon(R) Silver 4116 CPU @ 2.10GHz
BenchmarkHashicorpSetGet
BenchmarkHashicorpSetGet-8 6939919 762.1 ns/op 1 B/op 0 allocs/op
BenchmarkCloudflareSetGet
BenchmarkCloudflareSetGet-8 47835366 141.1 ns/op 16 B/op 1 allocs/op
BenchmarkEcacheSetGet
BenchmarkEcacheSetGet-8 64053291 94.14 ns/op 1 B/op 0 allocs/op
BenchmarkLxzanSetGet
BenchmarkLxzanSetGet-8 54651466 115.8 ns/op 0 B/op 0 allocs/op
BenchmarkFreelruSetGet
BenchmarkFreelruSetGet-8 46037301 130.8 ns/op 0 B/op 0 allocs/op
BenchmarkPhusluSetGet
BenchmarkPhusluSetGet-8 67405720 91.43 ns/op 0 B/op 0 allocs/op
BenchmarkNoTTLSetGet
BenchmarkNoTTLSetGet-8 67561371 85.05 ns/op 0 B/op 0 allocs/op
BenchmarkCcacheSetGet
BenchmarkCcacheSetGet-8 12946960 499.0 ns/op 25 B/op 2 allocs/op
BenchmarkRistrettoSetGet
BenchmarkRistrettoSetGet-8 58647771 103.5 ns/op 23 B/op 1 allocs/op
BenchmarkTheineSetGet
BenchmarkTheineSetGet-8 15404998 381.8 ns/op 2 B/op 0 allocs/op
BenchmarkOtterSetGet
BenchmarkOtterSetGet-8 70728278 88.09 ns/op 4 B/op 0 allocs/op
PASS
ok command-line-arguments 117.662s
The GC scan times as below. Check github gcscan action for more results and details.
GC scan times on keysize=16(string), valuesize=8(int), cachesize in (100000,200000,400000,1000000)
// env GODEBUG=gctrace=1 go run gcscan.go phuslu 1000000
package main
import (
"fmt"
"os"
"runtime"
"runtime/debug"
"strconv"
"time"
theine "github.com/Yiling-J/theine-go"
"github.com/cespare/xxhash/v2"
cloudflare "github.com/cloudflare/golibs/lrucache"
ristretto "github.com/dgraph-io/ristretto/v2"
freelru "github.com/elastic/go-freelru"
hashicorp "github.com/hashicorp/golang-lru/v2/expirable"
ccache "github.com/karlseguin/ccache/v3"
lxzan "github.com/lxzan/memorycache"
otter "github.com/maypok86/otter"
ecache "github.com/orca-zhang/ecache"
phuslu "github.com/phuslu/lru"
)
const keysize = 16
var repeat, _ = strconv.Atoi(os.Getenv("repeat"))
var keys []string
func main() {
name := os.Args[1]
cachesize, _ := strconv.Atoi(os.Args[2])
keys = make([]string, cachesize)
for i := range cachesize {
keys[i] = fmt.Sprintf(fmt.Sprintf("%%0%dd", keysize), i)
}
map[string]func(int){
"nottl": SetupNottl,
"phuslu": SetupPhuslu,
"freelru": SetupFreelru,
"ristretto": SetupRistretto,
"otter": SetupOtter,
"lxzan": SetupLxzan,
"ecache": SetupEcache,
"cloudflare": SetupCloudflare,
"ccache": SetupCcache,
"hashicorp": SetupHashicorp,
"theine": SetupTheine,
}[name](cachesize)
}
func SetupNottl(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache := phuslu.NewLRUCache[string, int](cachesize)
runtime.GC()
for range repeat {
for i := range cachesize {
cache.Set(keys[i], i)
}
runtime.GC()
}
}
func SetupPhuslu(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache := phuslu.NewTTLCache[string, int](cachesize)
runtime.GC()
for range repeat {
for i := range cachesize {
cache.Set(keys[i], i, time.Hour)
}
runtime.GC()
}
}
func SetupFreelru(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache, _ := freelru.NewSharded[string, int](uint32(cachesize), func(s string) uint32 { return uint32(xxhash.Sum64String(s)) })
runtime.GC()
for range repeat {
for i := range cachesize {
cache.AddWithLifetime(keys[i], i, time.Hour)
}
runtime.GC()
}
}
func SetupOtter(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache, _ := otter.MustBuilder[string, int](cachesize).WithVariableTTL().Build()
runtime.GC()
for range repeat {
for i := range cachesize {
cache.Set(keys[i], i, time.Hour)
}
runtime.GC()
}
}
func SetupEcache(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache := ecache.NewLRUCache(1024, uint16(cachesize/1024), time.Hour)
runtime.GC()
for range repeat {
for i := range cachesize {
cache.Put(keys[i], i)
}
runtime.GC()
}
}
func SetupRistretto(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache, _ := ristretto.NewCache(&ristretto.Config[string, int]{
NumCounters: int64(10 * cachesize), // number of keys to track frequency of (10M).
MaxCost: int64(cachesize), // maximum cost of cache (1M).
BufferItems: 64, // number of keys per Get buffer.
})
runtime.GC()
for range repeat {
for i := range cachesize {
cache.SetWithTTL(keys[i], i, 1, time.Hour)
}
runtime.GC()
}
}
func SetupLxzan(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache := lxzan.New[string, int](
lxzan.WithBucketNum(128),
lxzan.WithBucketSize(cachesize/128, cachesize/128),
lxzan.WithInterval(time.Hour, time.Hour),
)
runtime.GC()
for range repeat {
for i := range cachesize {
cache.Set(keys[i], i, time.Hour)
}
runtime.GC()
}
}
func SetupTheine(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache, _ := theine.NewBuilder[string, int](int64(cachesize)).Build()
runtime.GC()
for range repeat {
for i := range cachesize {
cache.SetWithTTL(keys[i], i, 1, time.Hour)
}
runtime.GC()
}
}
func SetupCloudflare(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache := cloudflare.NewMultiLRUCache(1024, uint(cachesize/1024))
runtime.GC()
for range repeat {
for i := range cachesize {
cache.Set(keys[i], i, time.Now().Add(time.Hour))
}
runtime.GC()
}
}
func SetupCcache(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache := ccache.New(ccache.Configure[int]().MaxSize(int64(cachesize)).ItemsToPrune(100))
runtime.GC()
for range repeat {
for i := range cachesize {
cache.Set(keys[i], i, time.Hour)
}
runtime.GC()
}
}
func SetupHashicorp(cachesize int) {
defer debug.SetGCPercent(debug.SetGCPercent(-1))
cache := hashicorp.NewLRU[string, int](cachesize, nil, time.Hour)
runtime.GC()
for range repeat {
for i := range cachesize {
cache.Add(keys[i], i)
}
runtime.GC()
}
}| GCScan | 100000 | 200000 | 400000 | 1000000 |
|---|---|---|---|---|
| nottl | 1 ms | 3 ms | 6 ms | 15 ms |
| phuslu | 1 ms | 3 ms | 6 ms | 14 ms |
| ristretto | 2 ms | 4 ms | 7 ms | 13 ms |
| freelru | 2 ms | 4 ms | 7 ms | 16 ms |
| lxzan | 2 ms | 5 ms | 8 ms | 19 ms |
| cloudflare | 5 ms | 11 ms | 21 ms | 56 ms |
| otter | 5 ms | 12 ms | 22 ms | 58 ms |
| ecache | 5 ms | 11 ms | 23 ms | 60 ms |
| ccache | 5 ms | 11 ms | 24 ms | 60 ms |
| hashicorp | 9 ms | 17 ms | 35 ms | 79 ms |
| theine | 7 ms | 15 ms | 35 ms | 83 ms |
The Memory usage result as below. Check github memory action for more results and details.
memory usage on keysize=16(string), valuesize=8(int), cachesize in (100000,200000,400000,1000000,2000000,4000000)
// memusage.go
package main
import (
"fmt"
"os"
"runtime"
"time"
"strconv"
theine "github.com/Yiling-J/theine-go"
"github.com/cespare/xxhash/v2"
cloudflare "github.com/cloudflare/golibs/lrucache"
ristretto "github.com/dgraph-io/ristretto/v2"
freelru "github.com/elastic/go-freelru"
hashicorp "github.com/hashicorp/golang-lru/v2/expirable"
ccache "github.com/karlseguin/ccache/v3"
lxzan "github.com/lxzan/memorycache"
otter "github.com/maypok86/otter"
ecache "github.com/orca-zhang/ecache"
phuslu "github.com/phuslu/lru"
)
const keysize = 16
var keys []string
func main() {
name := os.Args[1]
cachesize, _ := strconv.Atoi(os.Args[2])
keys = make([]string, cachesize)
for i := range cachesize {
keys[i] = fmt.Sprintf(fmt.Sprintf("%%0%dd", keysize), i)
}
var o runtime.MemStats
runtime.ReadMemStats(&o)
map[string]func(int){
"nottl": SetupNottl,
"phuslu": SetupPhuslu,
"freelru": SetupFreelru,
"ristretto": SetupRistretto,
"otter": SetupOtter,
"lxzan": SetupLxzan,
"ecache": SetupEcache,
"cloudflare": SetupCloudflare,
"ccache": SetupCcache,
"hashicorp": SetupHashicorp,
"theine": SetupTheine,
}[name](cachesize)
var m runtime.MemStats
runtime.ReadMemStats(&m)
fmt.Printf("%s\t%d\t%v MB\t%v MB\t%v MB\n",
name,
cachesize,
(m.Alloc-o.Alloc)/1048576,
(m.TotalAlloc-o.TotalAlloc)/1048576,
(m.Sys-o.Sys)/1048576,
)
}
func SetupNottl(cachesize int) {
cache := phuslu.NewLRUCache[string, int](cachesize)
for i := range cachesize {
cache.Set(keys[i], i)
}
}
func SetupPhuslu(cachesize int) {
cache := phuslu.NewTTLCache[string, int](cachesize)
for i := range cachesize {
cache.Set(keys[i], i, time.Hour)
}
}
func SetupFreelru(cachesize int) {
cache, _ := freelru.NewSharded[string, int](uint32(cachesize), func(s string) uint32 { return uint32(xxhash.Sum64String(s)) })
for i := range cachesize {
cache.AddWithLifetime(keys[i], i, time.Hour)
}
}
func SetupOtter(cachesize int) {
cache, _ := otter.MustBuilder[string, int](cachesize).WithVariableTTL().Build()
for i := range cachesize {
cache.Set(keys[i], i, time.Hour)
}
}
func SetupEcache(cachesize int) {
cache := ecache.NewLRUCache(1024, uint16(cachesize/1024), time.Hour)
for i := range cachesize {
cache.Put(keys[i], i)
}
}
func SetupRistretto(cachesize int) {
cache, _ := ristretto.NewCache(&ristretto.Config[string, int]{
NumCounters: int64(10 * cachesize), // number of keys to track frequency of (10M).
MaxCost: int64(cachesize), // maximum cost of cache (1M).
BufferItems: 64, // number of keys per Get buffer.
})
for i := range cachesize {
cache.SetWithTTL(keys[i], i, 1, time.Hour)
}
}
func SetupLxzan(cachesize int) {
cache := lxzan.New[string, int](
lxzan.WithBucketNum(128),
lxzan.WithBucketSize(cachesize/128, cachesize/128),
lxzan.WithInterval(time.Hour, time.Hour),
)
for i := range cachesize {
cache.Set(keys[i], i, time.Hour)
}
}
func SetupTheine(cachesize int) {
cache, _ := theine.NewBuilder[string, int](int64(cachesize)).Build()
for i := range cachesize {
cache.SetWithTTL(keys[i], i, 1, time.Hour)
}
}
func SetupCloudflare(cachesize int) {
cache := cloudflare.NewMultiLRUCache(1024, uint(cachesize/1024))
for i := range cachesize {
cache.Set(keys[i], i, time.Now().Add(time.Hour))
}
}
func SetupCcache(cachesize int) {
cache := ccache.New(ccache.Configure[int]().MaxSize(int64(cachesize)).ItemsToPrune(100))
for i := range cachesize {
cache.Set(keys[i], i, time.Hour)
}
}
func SetupHashicorp(cachesize int) {
cache := hashicorp.NewLRU[string, int](cachesize, nil, time.Hour)
for i := range cachesize {
cache.Add(keys[i], i)
}
}| 100000 | 200000 | 400000 | 1000000 | 2000000 | 4000000 | |
|---|---|---|---|---|---|---|
| nottl | 3 MB | 6 MB | 13 MB | 39 MB | 77 MB | 155 MB |
| phuslu | 4 MB | 8 MB | 16 MB | 46 MB | 92 MB | 185 MB |
| ristretto | 13 MB | 12 MB | 28 MB | 58 MB | 150 MB | 291 MB |
| lxzan | 8 MB | 17 MB | 35 MB | 95 MB | 191 MB | 379 MB |
| otter | 13 MB | 27 MB | 54 MB | 104 MB | 209 MB | 418 MB |
| freelru | 6 MB | 13 MB | 27 MB | 112 MB | 224 MB | 448 MB |
| ecache | 11 MB | 22 MB | 44 MB | 123 MB | 238 MB | 468 MB |
| theine | 15 MB | 31 MB | 62 MB | 178 MB | 357 MB | 714 MB |
| cloudflare | 16 MB | 33 MB | 64 MB | 183 MB | 358 MB | 716 MB |
| ccache | 16 MB | 32 MB | 65 MB | 182 MB | 365 MB | 730 MB |
| hashicorp | 18 MB | 37 MB | 57 MB | 241 MB | 484 MB | 967 MB |
- nottl saves 20% memory usage compared to phuslu by removing its ttl functionality, resulting in a slight increase in throughput.
- ristretto employs a questionable usage pattern due to its rejection of items via a bloom filter, resulting in a lower hit ratio.
- freelru overcommits the cache size to the next power of 2, leading to higher memory usage particularly at larger cache sizes.
It is a classic sharded LRU implementation, so the hit ratio is comparable to or slightly lower than a regular LRU.
LRU is licensed under the MIT License. See the LICENSE file for details.
For inquiries or support, contact phus.lu@gmail.com or raise github issues.