Add some perf tests to compare Mojo overhead.

mojo/edk/test/run_all_perftests.cc

Index: mojo/edk/test/run_all_perftests.cc
diff --git a/mojo/edk/test/run_all_perftests.cc b/mojo/edk/test/run_all_perftests.cc
index 50c41d7a758123999747f94e77bef48bd5170bbd..f426a2a5b199843607aeb1bebfb2696ee26b168b 100644
--- a/mojo/edk/test/run_all_perftests.cc
+++ b/mojo/edk/test/run_all_perftests.cc
@@ -21,10 +21,7 @@ int main(int argc, char** argv) {
 
   mojo::edk::Init();
   base::TestIOThread test_io_thread(base::TestIOThread::kAutoStart);
-  // Leak this because its destructor calls mojo::edk::ShutdownIPCSupport which
-  // really does nothing in the new EDK but does depend on the current message
-  // loop, which is destructed inside base::LaunchUnitTests.
-  new mojo::edk::test::ScopedIPCSupport(test_io_thread.task_runner());
+  mojo::edk::test::ScopedIPCSupport ipc_support(test_io_thread.task_runner());
   mojo::test::TestSupport::Init(new mojo::edk::test::TestSupportImpl());
 
   return test.Run();