I was on the road the last few days counseling our clients on how performance issues can be detected in .NET applications. Upon perusing my last entry once more, I realized that my account was not complete, since I haven’t yet informed you about the new features SpeedTrace 3.3 is going to bring to the market soon. Therefore, I’d like to just provide you with a short outline of some of the coolest features this tool has.
SpeedTrace 3.3 features:
- Tracing of multiple processes:
You can run several instances of SpeedTrace in parallel, so you can profile/trace different processes simultaneously in order to help you to focus on certain segments of your application at the same time. - Call history scoping:
SpeedTrace enables you to open partial traces just to see the callees of a given function by focusing exclusively on the calls made within a given segment. - Generating trace output from a terminated process:
An application that terminates unexpectedly may cause a real headache and waste a lot of time if you cannot analyze its cause properly. Due to SpeedTrace’s robust architecture, it yields reliable results thorough to the last drop even when your application crashes. - Hotspot filter:
Whenever you activate it, SpeedTrace will only display the nodes with own times higher than the specified threshold. - Comparing profile results:
It is important to know whether changes made within your code have a positive upshot or predominantly negative side effects. You can compare profile results with counterpoint method timings and allocation issues. - Triggers: Specify the method that will start or stop the profiling process in order to analyze specific behavior within specified boundaries.
- Black box filter:
3rd party components that cannot be improved always flood the analysis with irrelevant information. SpeedTrace, however, can encapsulate these components by just providing transitions for the ones you need. - Data trace:
SpeedTrace helps you to locate bugs (e.g. programming errors, deadlocks, missing or wrong version of DLL) easily, allowing you to examine the data flow channeled through the various functions. - Code location in Microsoft Visual Studio:
Whenever you find an issue, a quick tune of the code is required. You can always click on the Source view in SpeedTrace to locate your code in Visual Studio.
