New powerful tools and components

In Version 4.5 we provide a set of new specialized tools and components to aid developers throughout the product lifecycle:

• dot.NET Runtime-Analyzer : This separately available tool is geared to optimize scenario startup times for best performance and ultimate user satisfaction
• Decompiler: The integrated decompiler component allows you to reconstruct implementation details even in the case of missing source code

New powerful features

Version 4.5 features added functionality to increase diagnostics productivity and simplify navigation for high-yield analysis:

• Caller/Callee Backtraces for simplified load trigger investigation
• What-IF-Performance-Forecasts to create and compare scenario simulations
• Context-sensitive Call Stack History
• Enhanced analytics: percentiles, clustering, and more …

Enhanced user experience and optimization

Several aspects have been highly optimized to enhance user experience:

• Syntax Highlighting
• CPU Usage-View – ThreadStack Viewer to view thread stacks at particular times with a mouse click plus reduced memory consumption
• Enhanced Grouping/Color-Definitions/New LayerDefintion
• Bug fixes and runtime optimization

Outlook for 2013

We are going to release our current Version 4.5 in May next year. In the meantime, we’re also working on a .NET Memory Profiler, which will be also included in SpeedTrace Pro.

Seasonal greetings

The SpeedTrace Pro Team would like to thank you all for your appreciation, support and valuable feedback and warmly welcomes you to continue making good use of our products and support.

We wish you all a Merry Christmas and a Happy New Year!

Your SpeedTrace Pro Team

Alois Kraus, a .Net enthusiast, performance expert and interesting blog writer recently reviewed SpeedTrace Pro in his blog the other day. Since it poignantly displays SpeedTrace Pro’s competitive edge, we would like to share it with you: geekswithblogs.net

SpeedTrace Pro tutorial

Jon Shoeberg, senior architect, software engineer and consultant, even wrote a Youtube tutorial elegantly showing how easily you can sift out conclusive results for bug-trapping and improving app performance:

We would like express our cordial thanks and welcome people to share their experience with our product.

The tool SpeedTrace JITAnalyzer was renamed to RuntimeAnalyzer due to its extended scope:

- Startup-Optimizer: Regarding the base functionality, please refer to the following link

- Extended base functionality

• Attach to upcoming & already running processes: Attach to processes
  which are started outside the RuntimeAnalyzer tool
• Open/Save Snapshot: The open operation views saved snapshots,
  the save operation stores the captured runtime information

- Added Exception-Hunter: The built-in exception hunter shows all exceptions including time stamps, exception details and thread stacks. Each stack item contains line-level information to display the source code. In the following example the RuntimeAnalyzer pinpoints the line where a divide by zero exception was raised.

- Added ThreadStack-Analyzer: The thread stacks shows all running threads, its cpu time consumption and current call stacks at a specified sampling rate. The thread stack view allows you to examine the thread stacks at regular times (sampling) to investigate application hangups or any other simple performance issues.

- Added Decompiler-functionality to the source code preview (C#/IL): In case of missing source codes or debug info, methods of interested can be easily decompiled to either C# or IL.

How do we optimize startup performance?

Besides eliminating causes of bottlenecks induced by poor software algorithms affecting application startup, we need to also provide the best infrastructure for our software to run. The key task then is to provide proper software deployment. Before .Net software can run on the CPU, .Net assemblies must be translated from IL to native code. So we should deploy native images for assemblies which cause high time losses in order to minimize annoying runtime overheads.

How do we find out where time losses were triggered from?

SpeedTrace Pro introduces a new tool, the JITAnalyzer, to assist you in your investigation. It pinpoints how often and where JIT compilations occur and how much time is wasted due to current software deployment. Furthermore, it provides relocation information to help you reduce memory footprints of enterprise applications.

If you are interested in testing a preview version of the JITAnalyzer tool, please feel free to contact us:

Request SpeedTrace Pro → 10-day free trial

We are happy to announce the final release of SpeedTrace Pro’s new Version 4.0.

Request your trial license here: http://speedtracepro.com/support-download/download/

Amplified tracing scope now includes unmanaged code!

One of SpeedTrace Pro’s Version 4.0 key features this year is a new tracing method especially designed to embrace non-dotNET code. This newly added piece of cutting-edge functionality enables you to trace also non-dotNET platforms such as VBA, JavaScript, unmanaged C++, etc. all in one go. The advantage of this is precious, since a lot of .Net applications consist of mixed code nowadays.

Layer Breakdown – the ideal management reporting tool

The new dashboard ‘Layer Breakdown’ now provides you with a quick bird’s eye overview of the application’s overall time consumption layer by layer, so you can see at one glance where the most time was consumed. It is the ideal management reporting and controlling tool to help testers, consultants, architects and managers to scan the critical areas and to check what still needs to be done …. Besides, the layers can be customized, so that you are able to view your trace output from all kinds of different perspectives.

Powerful new features and extended functionality

• CPU time tracking (tracing, sampling) offers quick CPU usage checks enabling you to zoom from summary views down to specific time frames and even functions within the same scenario
• Splitting and merging multiple TraceFiles allows you to customize your tracing activity (ideal when working in team and task force constellations on large apps)

Enhancements on all levels

• Easier navigation / IIS tracing with a whole set of new browsers (project, application, website) all adding up to a fabulously comfortable UI feeling
• Even faster tracing due to structural improvements and enhancements to code and implementation

Outlook for next CTP

We are already in our starting blocks to tackle the next version’s challenges to add further leading-edge functionality and possibly widen our portfolio.

Welcome to Download

Send us your feedback. Comments and suggestions appreciated.
SpeedTrace Pro New Version 4.0 – ready for download.

Try SpeedTrace Pro → 10-day free trial

SpeedTrace Pro – our new 4.0 Version Release!

We are proud to announce our new release featuring a whole new range of added functionality highlights such as:

• Unmanaged Code tracing
• Layer Breakdown
• CPU tracking
• … and more …

A more detailed entry on the new SpeedTrace functionality is to follow up briefly!

SpeedTrace Pro – our new website!

We have moved our website from perfinity.de to per.auriga.uberspace.de/htdocs and created a whole new design presentation that lives up to our .NET trace profiler product. So SpeedTrace Pro now has its own homepage that serves to improve communication on our profiler solution and aptly portray our expertise in the field of software performance management.

We hope you like it. We are still making improvements, so feedback and suggestions are warmly appreciated.

Welcome to http://speedtracepro.com!

Performance consulting – our new service!

Finally, we are happy to let you know that our SpeedTrace Pro team now has a consulting division that offers consulting in many forms:

• Trouble-shooting
• Performance analysis
• Online/ In-house training

Feel free to contact us!

New in Version 4.0.32:

Enhancements:

• supports ReSharper 6.1 integration
• improved trace session manager

Bug fixes:

• fix: improved profiler overhead calculation
• fix: thread assignment

Try the .NET Tracer and Profiler SpeedTrace Pro 4.0.32 today (available as 32-bit and 64-bit version).

Request a 10-day free trial license →

If you need software test automation, you need SpeedTrace Pro!

The SpeedTrace Pro Team would like to thank you all for your appreciation, support and valuable feedback.

We wish you all a Merry Christmas and a Happy New Year!

Your SpeedTrace Pro Team

Version 4.0.23 brings enhancements and some bugfixes.

These are the changes since version 4.0.19:

Enhancements

• Improved stop trace performance
• Using rtf style for call history
• Improved performance for workspace export
• Added warning in case of missing permissions for certain typed
• Callhistory – changed DrillDownTrace. OpenScope – clipped timerange of the shadow stack parents to the selected time range
• Callhistory – added drop down list to select trace session file
• Richformat-Text for trace find results & bookmarks
• Optimized simplify filters performance
• Trace Analyzer UI: Improved User Experience

Fixed Bugs

• fix Activate/Deactivate button dimming
• fix: Multiple Call History – corrected chart legend
• fix: trace engine cyclic trigger when focus on start trigger thread is activated
• fix: improved workspace handling for path too long exceptions
• fix: CallHistory View – keeps selected Time range in case of resize events
• fix: Trace agent engine/trace processing – trigger corrections (stop parent, cyclic)
• fix: Client dashboard – copy & paste shortcuts

Try SpeedTrace Pro now → 10-day free trial

SpeedTrace Pro is not just a one night affair. It is really worth getting involved in a long-lasting relationship with all its responsibilities and refreshing surprises. To show this we have created a more elaborate web scenario (below) to demonstrate how SpeedTrace masters the complexities of web-based applications with great ease and elegance.

We are going to examine the client server communication in a WCF RIA Services scenario where the web server is IIS 7 running ASP.Net 4 and the client is Firefox running Silverlight 4.

Trace project setup

In order to analyze the communication between client and server, we need to trace the Silverlight client and the IIS server at the same time. So let’s first create the necessary trace project: We will choose a slightly modified version of an IIS project because we want to trace both the IIS server and the Firefox plugin container. Since we want to perform local tracing, we simply keep the local host as target computer (#1). After having specified the target website (#2) which is going to be traced we add the Firefox plugin container to the w3wp (#3) worker process as another application to be traced and inform SpeedTrace to trace exactly two processes (#4). If we have multiple w3wps running, we can filter the application pool or also trace multiple instances. Finally, we uncheck start recording immediately, because we do not want to trace the application startup.

Trace project activation

After having activated the trace project, the default browser will open the specified website, so that the IIS webserver and the Silverlight application get activated and the SpeedTrace trace agent will attach to the processes.

The trace session manager will display information about the active trace sessions (CPU usage, target process name, target computer, trace state etc.) and can also be used to control the trace state.

Trace scenario

The Silverlight client displays data from the Northwind Database which is hosted by the MS SQL Server:

Our trace scenario is to perform paging within the data grid. Whenever we reach the last page, we re-assort the data and page again until we have executed 32 paging operations. Tracing will be started shortly before the first and stopped after the last paging operation.

Trace Results

As soon as we stop the trace, we will receive the trace output files for the affected trace sessions. In our case, one trace file for the IIS server, and one for the Silverlight client.

Trace Analysis

Now, here comes the interesting part – the trace analysis. Using SpeedTrace it is easy to find the reason for performance bottlenecks, but what about analyzing communication times between two processes?

Let’s first merge the two trace files into one analysis session:

Get the Big Picture – Layer Breakdown over time

The layer breakdown, the bird’s eye view, would look like this for the Silverlight client
(showing high presentation layer consumption):

And like this for the IIS server (showing high DAL consumption):

CPU usage – Timeline

Finally, before we get started with the actual analysis, let’s have a quick look at the CPU time usage of both processes.

For the IIS server:

For the Silverlight client:

Communication Analysis

Let’s first talk about the involved APIs:

Having analyzed the client and server trace, you will see that …

… the Silverlight client …

1) … starts service requests via ServiceChannelProxy.InvokeBeginService()
2) … receives the response via HttpChannelFactory.OnGetResponse()
3) … and ends it via ServiceChannelProxy.InvokeEndService()

… the IIS server will receive and dispatch the request within ChannelHandler.HandleRequest().

So let’s have a look at the timeline regarding the mentioned APIs: We do this by using Call History …

When you look at the times, you will notice that

#1 The request is sent to the server in just 3 ms …
#2 After a server processing time of about 22 ms …
#3.1 … the client is notified in just 3 ms about the response …
#3.2 It takes another 5 ms to get the response … and
#4 … another 16 ms (!) to finalize the query

Why is that (#3.2, #4), why do we lose around 20 ms?

Let’s drill down and open the trace for #3.2

Apparently the time was wasted for thread synchronization. Fine, then let’s analyze #5:

Another deferred call? This time we have lost 16 ms, so 21 ms in total because of deferred calls.

Performance summary:

Conclusion

SpeedTrace Pro is a powerful tool – much more than a simple profiler, it helps you to investigate complicated issues, find their root causes, and design the right solutions. This, of course, could not have been done without the sophisticated functionality it provides.

Due to its power and reliability, SpeedTrace Pro is no tool for just for one night. If you are interested in a long-term relationship being proactively responsible for your Continuous Application Performance Management, please feel free to just fall in love with it.