Add StateT benchmark.

bench/src/main/scala/cats/bench/StateTBench.scala

@@ -0,0 +1,51 @@
+package cats.bench
+
+import cats.{Eval, Id, Monad}
+import cats.data.StateT
+import org.openjdk.jmh.annotations.{Benchmark, Scope, State}
+
+@State(Scope.Benchmark)
+class StateTBench {
+
+  def iteratedFlatMap[F[_]](n: Int)(implicit F: Monad[F]): StateT[F, Int, Int] =
+    (1 to n).foldLeft(StateT.pure[F, Int, Int](n))((st, _) => st.flatMap(i => StateT(_ => F.pure((i-1, i-1)))))
+
+  def recursiveFlatMap[F[_]](implicit F: Monad[F]): StateT[F, Int, Int] =
+    StateT[F, Int, Int](n => F.pure((n-1, n-1))).flatMap(i =>
+        if(i > 0) recursiveFlatMap
+        else StateT.pure[F, Int, Int](0)
+    )
+
+  @Benchmark
+  def strictIteratedFlatMap100: Int =
+    iteratedFlatMap[Id](100).runA(0)
+
+  @Benchmark
+  def strictIteratedFlatMap100k: Int =
+    iteratedFlatMap[Id](100000).runA(0)
+
+  @Benchmark
+  def strictRecursiveFlatMap100: Int =
+    recursiveFlatMap[Id].runA(100)
+
+  @Benchmark
+  def strictRecursiveFlatMap100k: Int =
+    recursiveFlatMap[Id].runA(100000)
+
+  @Benchmark
+  def trampolinedIteratedFlatMap100: Int =
+    iteratedFlatMap[Eval](100).runA(0).value
+
+  @Benchmark
+  def trampolinedIteratedFlatMap100k: Int =
+    iteratedFlatMap[Eval](100000).runA(0).value
+
+  @Benchmark
+  def trampolinedRecursiveFlatMap100: Int =
+    recursiveFlatMap[Eval].runA(100).value
+
+  @Benchmark
+  def trampolinedRecursiveFlatMap100k: Int =
+    recursiveFlatMap[Eval].runA(100000).value
+
+}