Identifying Equivalent Objects to Reduce Memory Consumption

Executing an application may trigger the creation of a large amount of objects. For many applications, a large portion of these objects are unnecessary and their creation could simply be avoided. We describe a lightweight profiling technique to identity “equivalent” objects. Such equivalent objects are simply redundant and may be shared or reused to reduce the memory footprint. We propose object-centric execution blueprint, a visual representation to help practitioners identify cases where objects may be reused instead of being redundant.