Optimised implementation of List.filter/filterNot

1 files changed, 57 insertions, 0 deletions

diff --git a/test/benchmarks/src/main/scala/scala/collection/immutable/ListBenchmark.scala b/test/benchmarks/src/main/scala/scala/collection/immutable/ListBenchmark.scala
new file mode 100644
index 0000000000..5ed30de7f6
--- /dev/null
+++ b/test/benchmarks/src/main/scala/scala/collection/immutable/ListBenchmark.scala
@@ -0,0 +1,57 @@
+package scala.collection.immutable
+
+import java.util.concurrent.TimeUnit
+
+import org.openjdk.jmh.annotations._
+
+object ListBenchmark {
+  case class Content(value: Int)
+}
+
+@BenchmarkMode(Array(Mode.AverageTime))
+@Fork(2)
+@Threads(1)
+@Warmup(iterations = 10)
+@Measurement(iterations = 10)
+@OutputTimeUnit(TimeUnit.NANOSECONDS)
+@State(Scope.Benchmark)
+class ListBenchmark {
+  import ListBenchmark._
+  @Param(Array("0", "1", "10", "100", "1000"))
+  var size: Int = _
+
+  var values: List[Content] = _
+  var mid: Content = _
+  var last: Content = _
+
+
+  @Setup(Level.Trial) def initKeys(): Unit = {
+    values = List.tabulate(size)(v => Content(v))
+    mid = Content(size / 2)
+    last = Content(Math.max(0,size -1))
+  }
+
+  @Benchmark def filter_includeAll: Any = {
+    values.filter(v => true)
+  }
+
+  @Benchmark def filter_excludeAll: Any = {
+    values.filter(_ => false)
+  }
+
+  @Benchmark def filter_exc_mid: Any = {
+    values.filter(v => v.value != mid.value)
+  }
+
+  @Benchmark def filter_from_mid: Any = {
+    values.filter(v => v.value <= mid.value)
+  }
+
+  @Benchmark def filter_exc_last: Any = {
+    values.filter(v => v.value != last.value)
+  }
+
+  @Benchmark def filter_only_last: Any = {
+    values.filter(v => v.value == last.value)
+  }
+}