from $395
($494 including support & upgrades)
-
- Purchase this now
-
-
ANTS Profiler™
Walk-through 1: Performance profiling
This walk-through will show you:
- How to set up a new project and start profiling
- How to interact with the timeline
- How to drill down into the results to identify the bottleneck
- How to extract the most out of the source-code pane
Step 8: Exploring the source-code pane
We have selected the Mandelbrot.Algorithm.EvaluateUsingComplexNumbers method on the call tree and we are presented with the source-code file in the source-code pane [1st screenshot]. There, we can obtain quantitative data such as hit count and line-level timings, so we do not have to waste time guessing where bottlenecks may be. Expensive lines of code are highlighted with a red bar and indexed on the heat map bar on the right-hand side for quick navigation.
Line 62 is flagged as an expensive line. We check the time spent in seconds, and we find that it took 9 seconds to execute [2nd screenshot]. We know Mandelbrot well, and we know that this particular line of code should not take that long to run. This shows us that we have implemented an inefficient algorithm.
Note: The source-code pane is interactive and you can navigate it by clicking directly on a method call in order to jump to the source code file for that method. This would have been another easy way to get to the bottom of our problem.