Transplant jmh test OneShot.java from Loom

[Jong111]

Main changes

1. Create a standalone Maven project to meet the needs of JMH, including a pom.xml file, a Java class file, and a txt file which contains the result of the OneShot test
2. Remove the import of jdk.internal.vm.Continuation and jdk.internal.vm.ContinuationScope, which is no need in the test

Detailed diff

-import jdk.internal.vm.Continuation;
-import jdk.internal.vm.ContinuationScope;
-
 import java.util.concurrent.TimeUnit;

 import org.openjdk.jmh.annotations.*;
@@ -210,15 +207,22 @@ public class OneShot {
     }

     /**
-     * Creates and runs a continuation that yields at a given stack depth.
+     This benchmark will continuously create new continuations, which in KonaFiber means that
+     the program will keep requesting native memory and eventually run out of native memory,
+     so the benchmark will get a runtime error in KonaFiber. While running the benchmark with Loom
+     does not have this problem, because Loom creates the continuation in the JVM heap space,
+     and the stacked continuation will be cleaned up by the GC mechanism.
      */
-    @Benchmark
-    public void yield() {
-        Continuation cont = Yielder.continuation(paramCount, stackDepth, true);
-        cont.run();
-        if (cont.isDone())
-            throw new RuntimeException("continuation done???");
-    }
+//    /**
+//     * Creates and runs a continuation that yields at a given stack depth.
+//     */
+//    @Benchmark
+//    public void yield() {
+//        Continuation cont = Yielder.continuation(paramCount, stackDepth, true);
+//        cont.run();
+//        if (cont.isDone())
+//            throw new RuntimeException("continuation done???");
+//    }

[johnshajiang]

@Jong111 You may also want to provide a README.md for clarifying the way to build and run this test.

[Jong111]

@Jong111 You may also want to provide a README.md for clarifying the way to build and run this test. ok, i will add it

[miao-zheng]

1. 在readme里面没有看到测试结果，也没有loom的测试结果；
2. 没有在comment里看到，git diff列出当前文件和loom原始文件的差异

[Jong111]

1. 在readme里面没有看到测试结果，也没有loom的测试结果；
2. 没有在comment里看到，git diff列出当前文件和loom原始文件的差异

回复1：我将kona_fiber的测试结果放在了oneshot_test_result.txt中，您可以在该文件中查看，我需要将其转移到readme中吗？之后我把loom的测试结果也上传到指定文件； 回复2：感谢您的指出，之后我会上传Main.java和原始的OneShot.java之间的差异，请问是将git diff的结果作为注释保存在Main.java中还是单独新建一个文件保存？

[miao-zheng]

不需要新建一个文件，只要在comment里列出来就行

[Jong111]

不需要新建一个文件，只要在comment里列出来就行

好的

[miao-zheng]

看了git diff的文件，还是有很多非必要的改动，比如：

-                case 1: run1(maxDepth); break;
-                case 2: run2(maxDepth, new Arg()); break;
-                case 3: run3(maxDepth, new Arg(), new Arg()); break;
-                default: throw new Error("should not happen");
+                case 1:
+                    run1(maxDepth);
+                    break;
+                case 2:
+                    run2(maxDepth, new Arg());
+                    break;
+                case 3:
+                    run3(maxDepth, new Arg(), new Arg());
+                    break;
+                default:
+                    throw new Error("should not happen");
             }

建议这些非必要的改动全部去掉

[Jong111]

看了git diff的文件，还是有很多非必要的改动，比如：

-                case 1: run1(maxDepth); break;
-                case 2: run2(maxDepth, new Arg()); break;
-                case 3: run3(maxDepth, new Arg(), new Arg()); break;
-                default: throw new Error("should not happen");
+                case 1:
+                    run1(maxDepth);
+                    break;
+                case 2:
+                    run2(maxDepth, new Arg());
+                    break;
+                case 3:
+                    run3(maxDepth, new Arg(), new Arg());
+                    break;
+                default:
+                    throw new Error("should not happen");
             }

建议这些非必要的改动全部去掉

好的，已经改了