MicroBenchmark

Easy to use profiling tool to measure code performance in Unreal Blueprints or C++.

• Documentation
• C++ Doxygen Docs
• Support Forum
• Marketplace Link

Step-by-step usage guide

• Open the plugin settings of your project.
  
• Make sure MicroBenchmarkTimer is enabled. Restart UE eventually.
  
• Within any Blueprint Graph, right click and type benchmark to get the MicroBenchmarkSubsystem node.
  
• From the MicroBenchmarkSubsystem node, you’ll see all available functions under the Micro Benchmark tab. We’ll get into those in a minute.
  
• Optional: The plugin contains a demo blueprint actor containing the two example blueprints shown next. To open it, make sure you have Show Plugin Content enabled. Then look under Plugins/MicroBenchmarkTimer/DemoBlueprint or search for MicroBenchmarkDemoActor. Drag the blueprint into any level and open the blueprint graph.
  
• Tic-Toc Example. Let’s say under BeginPlay (1) of your level blueprint, you want to measure how long it takes to retrieve all StaticMeshActors (2). Simply add a Tic node (3) before the GetAllActorsOfClass node to start measuring. Add a Toc node (4) right after the GetAllActorsOfClass node to clock the time it took. To immediately see the timing result on the screen and in your log, enable the Show Log flag (5).
  
• OnTick Example. Now let’s say under EventTick (1) you’re doing a line trace (2) every frame. You could measure that as before, but that would result in too many log messages per second. Instead, give your Tic (3) a proper benchmark name (4) and disable the log flag in your Toc (5). Now have a look at the average timing across all line trace calls by calling the Print Micro Benchmark Results node (6). If you leave the benchmark name empty, all benchmark results will be printed. In this case, you’re printing the results with a B keyboard press. Though you could print results with a timer every 5 seconds or so, or on any other event you choose.
  
• That’s a wrap. Cheers.

FAQ

• Q: What happens if I call Tic again with the same benchmark name?
  A: The old Tic start-time gets overwritten and the execution time is measured from the last recent Tic call.
• Q: What happens if I call Toc again with the same benchmark name?
  A: Nothing.
• Q: Why is the runtime from Tic to Toc without any code inbeetween not 0?
  A: That’s the Blueprint overhead of calling those functions. On my machine it’s around 1.5 microseconds.
• Q: Why are the benchmark functions called Tic and Toc?
  A: The names are not inspired by a certain social network, but rather by two old-school MATLAB functions of the same name.

C++ Usage

• Enable the MicroBenchmark Plugin in you C++ project by including MicroBenchmarkTimer in your .Build.cs under PrivateDependencyModuleNames and make sure it is included and enabled in your .uproject file.
• Include MicroBenchmarkSubsystem.h in your .cpp
• Use the TIC and TOC macros (useable within any UObject) to measure the execution time of all the code inbetween.
• Full C++ API documentation here and in the source code itself.