Performance Benchmark Mistakes, Part Four

added by Grishma Govani
8/19/2013 2:12:08 PM


Last time in this series we talked about how forgetting to "charge" the expense of the jitter to your analysis can lead to inaccurate results: since the first run of a method incurs the cost of the jitter, you might want to leave that out of your average. Or you might want to leave it in! It depends on your scenario. But regardless, it is wise to be aware that there is a cost there, and that it needs to be taken into account when reasoning about measurements. The same is true of garbage collection, which is the subject of today's episode of this tutorial. So without further ado: