Try/Catch performance overhead

Benchmark created by Marco Rogers on


Description

Using try/catch inside of functions that allocate lots of variables introduces significant overhead. See this discussion related to node.js https://groups.google.com/forum/#!topic/nodejs-dev/E-Re9KDDo5w

Interestingly, other js engines don't have this overhead. All test runs have pretty uniform performance. But they are also significantly slower than the v8 control case.

Preparation HTML

<script>
  var tries = 10000;
  
  function test1() {
   function test(j) {
    var s = 0;
    for (var i = 0; i < j; i++) s = i;
    for (var i = 0; i < j; i++) s = i;
    for (var i = 0; i < j; i++) s = i;
    for (var i = 0; i < j; i++) s = i;
    return s;
   }
  
   test(tries);
  }
  
  function test2() {
   function test(j) {
    try {
     var s = 0;
     for (var i = 0; i < j; i++) s = i;
     for (var i = 0; i < j; i++) s = i;
     for (var i = 0; i < j; i++) s = i;
     for (var i = 0; i < j; i++) s = i;
     return s;
    }
    catch (ex) {}
   }
  
   test(tries);
  }
  
  function test3() {
   function test(j) {
    var s = 0;
    for (var i = 0; i < j; i++) s = i;
    for (var i = 0; i < j; i++) s = i;
    for (var i = 0; i < j; i++) s = i;
    for (var i = 0; i < j; i++) s = i;
    return s;
   }
  
   try {
    test(tries);
   }
   catch (ex) {}
  }
</script>

Test runner

Ready to run.

Testing in
TestOps/sec
control - no try/catch
test1();
ready
try/catch inside function
test2();
ready
try/catch outside function
test3();
ready

Revisions

You can edit these tests or add more tests to this page by appending /edit to the URL.