diff --git a/README.md b/README.md
index 48cecf4..6469519 100644
--- a/README.md
+++ b/README.md
@@ -171,7 +171,7 @@ There was only the test of `ants` Pool because my computer was crash when it rea
 
 **As you can see, `ants` can up to 2x~6x faster than goroutines without pool and the memory consumption is reduced by 10 to 20 times.**
 
-[1]: https://travis-ci.com/panjf2000/ants.svg?branch=master
+[1]: https://travis-ci.com/panjf2000/ants.svg?branch=develop
 [2]: https://travis-ci.com/panjf2000/ants
 [3]: https://godoc.org/github.com/panjf2000/ants?status.svg
 [4]: https://godoc.org/github.com/panjf2000/ants
diff --git a/ants_benchmark_test.go b/ants_benchmark_test.go
index 3668e0d..0c7728d 100644
--- a/ants_benchmark_test.go
+++ b/ants_benchmark_test.go
@@ -77,6 +77,7 @@ func BenchmarkGoroutineWithFunc(b *testing.B) {
 }
 
 func BenchmarkAntsPoolWithFunc(b *testing.B) {
+	defer ants.Release()
 	var wg sync.WaitGroup
 	p, _ := ants.NewPoolWithFunc(50000, func(i interface{}) error {
 		demoPoolFunc(i)
@@ -103,6 +104,8 @@ func BenchmarkGoroutine(b *testing.B) {
 
 func BenchmarkAntsPool(b *testing.B) {
 	p, _ := ants.NewPoolWithFunc(50000, demoPoolFunc)
+	p.Release()
+
 	b.ResetTimer()
 	for i := 0; i < b.N; i++ {
 		for j := 0; j < RunTimes; j++ {
diff --git a/ants_test.go b/ants_test.go
index 2334e42..0e1bc26 100644
--- a/ants_test.go
+++ b/ants_test.go
@@ -54,6 +54,7 @@ func TestDefaultPool(t *testing.T) {
 }
 
 func TestNoPool(t *testing.T) {
+	defer ants.Release()
 	var wg sync.WaitGroup
 	for i := 0; i < n; i++ {
 		wg.Add(1)