dot.NET Profiler

A few weeks ago we invited you to test your .NET Profiler & Tracer application.
With the test application SuspendTester, you had the opportunity to create an extreme multithreading test scenario.

SpeedTrace Pro has masted this test fast, easily and reliably.

We have received a lot of positive feedback, and some people followed our invitation to give video answers. Thanks a lot!

But see for yourself how SpeedTrace Pro left it’s major competitions way behind!

In our test lab we have repeated the test scenario 15 times and have come to the same results. For stability and reliability, you just can’t beat SpeedTrace Pro!

SpeedTrace Pro is a fully matured product which is able to resolve the root cause of problems ranging from general runtime to performance problems.

Stabilility/Reliability:
Check out the video comparison above and decide oneself.

Quality: Highly accurate results due to lowest profiling overhead of its fast trace engine (In the above videos you can see the runtime overtime raised by the various tools compared)

Traceability/Trace-Empowers Principle: SpeedTrace Pro records and stores detailed information of
application events for later reference with respect to causal analyses (Timestamps, transactions can be recorded even with data flow information) …
…to enable you to reconstruct what happened inside your application whenever
something goes wrong
…to resolve inexplicable application behavior (raised by exceptions, synchronization
issues, performance problems etc.)

Resolve runtime issues/Monitor problems: Due its low overhead and ‘traceability’-feature it can be used for monitoring applications

Remote/Distributed Tracing: SpeedTrace provides freely distributable trace agents, which can be used to record applications which are distributed throughout the network. In contrast to its competitors, SpeedTrace can trace multiple processes on any machines, and the trace results can be stored on a network share which is accessible from other clients analyzing the results.

Precise/Detailed Recording: Furthermore, SpeedTrace Pro allows you to record with greatest accuracy using the following unique recording features:

• Triggers define when to start and stop the recording of applications WITHOUT code
  modifications (invariably required when using other tools)
• Filters help you to define what to record to be able to track enterprise apps, breaking
  down the complexity by using features such as black-boxing
• DataTrace allows you to investigate the application data flow in case of problems by
  providing information that was passed to methods
• Callback API enables you to interactively react to application events by implementing a SpeedTrace Pro user extension without product code modifications.

Unique and powerful analysis features/Resolve Problems:

• Divide-and-conquer principle allows you to further break down the complexity by using features such as black-boxing (thus simplifying profiling info).
• Complex viewing options: By filtering information you can change your point of view
  when looking at the data, to empower yourself to follow the divide-and-conquer
  principle, for example, in order to get rid of implementation details of other teams,
  external components, etc. This allows development stakeholders to see only the
  information they are interested in.
• Get to the root: Powerful drill-downs enable you to work efficiently to find the real
  causes of problems

• Test Automation API/Prevent problems: SpeedTrace Pro provides you with a rich
  (recording/analysis) API to integrate SpeedTrace Pro into the application performance life cycle by regularly performing sanity checks during the development stage (-> automated testing).

• High quality support is provided by experienced performance consultants: Unlike other tools SpeedTrace Pro also offers consultancy services to resolve performance crises in major companies.

For more information view our documentation on our website, this dotnet-profiler-blog, visual studio gallery presentation.

Version 3.4.7 of our .Net
Tracing and Profiling Solution is available now.

A small bug in the DataTrace Engine was fixed.

Here you can download the latest build.

Common trace problems in virtual environments

In the wake of our consulting activities we encountered a minor hurdle while running SpeedTrace in a large-application virtual machine environment. Since some of our customers also came across similar problems I thought it might be useful to let you in on the discourse.

Unexpectedly high overhead behavior …

While running SpeedTrace in such an environment we were suddenly confronted with an unusual overhead of 4-6 times the application’s normal speed – quite atypical of SpeedTrace’s reliably unintrusive behavior. It was as if there was some kind of echoing or reverberating effect that was interfering with the tracing procedure.

… is often caused by the VMware

Our developers found the VMware architecture to be at the root of the problem. VMware happens to be the software commonly used in virtual machine environments. Besides, it looks like we are going to have to cope with VMware overhead for some time to come … In any case, the VM overhead would invariably occur with any other profiler even as good as SpeedTRace which has been repeatedly acclaimed for its unusual unintrusiveness, accuracy and speed.

SpeedTrace solution to VM yields excellent results!

We are happy to announce, however, that our SpeedTrace development team has devised a rather simple and elegant solution to the problem. We tried it out and it works perfectly down to the usual couple of nanoseconds SpeedTrace normally generates in terms of its extremely low overhead rates.

SpeedTrace provides 100% accurate
readouts with zero overhead!

In fact, with SpeedTrace you can even objectively calculate and thus eliminate the already minute overhead delays to yield maximum accuracy! The new 3.3.34 version about to go online (but can be sent immediately on request) even takes a step further to show these net results automatically.

Don’t forget to use blackboxing, triggers and filters!

SpeedTrace’s overwhelming accuracy, however, should not distract your attention from the variety of time- and resource-saving blackboxing, trigger, and filtering functions that Speedtrace offers to speed up your whole trace and profiling process. To profile efficiently, third-party software components, for example, can and should be isolated through blackboxing even before you make your first trace. SpeedTrace is not only incredibly accurate, it is also stunningly fast!

For best results with virtual machines please consult our SpeedTrace support team.