Remove bottlenecks from your code: ANTS Profiler 4 early access program

A few months ago we bit the bullet and decided to re-write ANTS Profiler, Red Gate's .net code profiling application. ANTS Profiler 3 was good, but AP4 is going to be remarkable. It's faster (between 5 and 30 times, depending on what you're doing); easier to use (we interviewed and videoed tens of users, and gathered data from hundreds, if not thousands, while designing AP4); and has a great timeline feature (you can highlight areas of high cpu and just see the code that was executed in those time slots).

We've now got a public build available. It's still a very early version but there's enough there for it to be useful. We're going to be releasing weekly builds from now on, so keep an eye out.

If you're writing .net code and want to find bottlenecks in your application, then either read Bart's post for some more information, or go straight to the messageboards to download the early build of ANTS Profiler 4.

Subscribe to the RSS feed

Comments

[Denis]

Perhaps you could have outlined how is this better than Visual Studio? Otherwise I do not feel compelled to learn about it.

[Magnus]

Visiting http://www.red-gate.com/messageboard/
gets me an HTTP 500 error message.

[James Moore]

Hi Magnus,

We experienced some problems with the server which hosts our message boards earlier, these should have now been resolved.

I'm sorry for any inconvenience caused.

James
--
James Moore
Red Gate Software Ltd

[Bart Read]

Hi Denis,

Thanks for posting. That's certainly a fair question, given that you've already handed over quite a chunk of cash for Visual Studio.

I think some of the main advantages of ANTS Performance Profiler 4 over the Visual Studio profiler are as follows (I'm basing this on the VS 2005 Team Edition profiler, which I have; I can't really speak for 2008):

1. It's generally much faster. It incurs *significantly* less overhead on your application as compared to running the VS profiler using instrumentation. Unlike VS we don't have a sampling mode, however our fastest instrumentation mode is only around 20 - 30% slower than sampling, and is obviously more accurate.

2. You can use AP4 with signed code. With VS you have to disable signing of your assemblies if you want to use instrumentation. Maybe not a big deal if you've only got one or two assemblies in your solution, but painful if you have ten, which is a pretty common scenario.

3. It's easier to use. I spent some time playing with the VS profiler this afternoon because I'd actually never used it before. I profiled Exception Hunter 1.1, so a real application, and not some toy example. I ran a direct comparison of AP4 and VS. What I discovered was that any time saving advantage offered by using the sampling mode was more than overshadowed by having to spend ages afterwards digging through the results trying to figure out what's going on, and bear in mind that I knew what I was looking for because I could already see the results in AP4. AP4 will show you the hottest stack trace through your application right at the top of your results, so you can instantly see where the bottleneck is.

4. AP4 provides line level timings. VS doesn't.

5. AP4 allows you to select the time period for which you want to view performance data. Let's say you have to perform a number of tasks *before* you get to the point where you can do the task where your performance bottleneck manifests itself. Do you really want the performance data from the previous tasks polluting the results for the task you're really interested in? I doubt it. With AP4 you can simply drag out over the time period you're interested in on the interactive timeline, and you'll see results only for that time period.

I think those are probably most of the major points. If you want to know more please take a look at my blog post:

http://www.simple-talk.com/community/blogs/bart/archive/2008/05/12/52623.aspx

Or just download the current build. Details are here:

http://www.red-gate.com/messageboard/viewforum.php?f=80

Obviously it's an early access build so it has some bugs in it (see my blog post, or look at the posts in the forum), but even taking this into account--and yes, I *am* totally biased--I hope you'll find the improvement is immediately clear.

Hope that's all helpful. And thanks again for posting.

Kind regards,
Bart