动态代理：如何处理嵌套方法调用(Dynamic Proxy: how to handle nested method calls)

 我会对@autoreleaspool发表评论，以帮助你解决问题。 
 
 Apple建议使用@autoreleaspool来关注内存。 以下段落取自Core Data 文档 ，但我相信也可以应用于这种情况： 
 
 
 
 与许多其他情况一样，当您使用Core Data导入数据文件时，重要的是要记住适用Cocoa应用程序开发的“常规规则”。 如果导入必须以某种方式解析的数据文件，则可能会创建大量临时对象。 这些会占用大量内存并导致分页。 就像使用非Core Data应用程序一样，您可以使用本地自动释放池块来限制内存中驻留的其他对象数量。 有关Core Data和内存管理之间交互的更多信息，请参阅“减少内存开销”。 
 
 
 基本上，@ autoreleasepool用作编译器在一旦超出范围时释放所有临时对象的提示。 您期望完全释放内存，这可能不是Apple框架的情况。 窗帘后面可能会有一些缓存（这只是想法）。 这就是剩下的1MB可能没问题的原因。 但是，为了安全起见，我建议增加迭代次数，看看会发生什么。 
 
 正如您在评论中提到的，您的循环很大并且是嵌套的，因此可能还会发生其他事情。 尝试摆脱所有额外的操作，看看会发生什么。 
 
 希望这有帮助，干杯！ 

I'll make my comment on @autoreleaspool a legitimate answer to help you out.
 
Apple suggests to use @autoreleaspool where the memory is in concern. The following paragraph is taken from Core Data documentation, but I believe can be applied in this situation as well:
 
 
 
In common with many other situations, when you use Core Data to import a data file it is important to remember “normal rules” of Cocoa application development apply. If you import a data file that you have to parse in some way, it is likely you will create a large number of temporary objects. These can take up a lot of memory and lead to paging. Just as you would with a non-Core Data application, you can use local autorelease pool blocks to put a bound on how many additional objects reside in memory. For more about the interaction between Core Data and memory management, see “Reducing Memory Overhead.”
 
 
Basically, @autoreleasepool serves as a hint for a compiler to release all temporary objects once they are out of bounds. You're expecting for the memory to be released completely, which might not be the case with Apple frameworks. There might be some caching going in behind the curtains (This is just and idea). That is why that remaining 1MB may be ok. However, just to be safe, I would recommend to increase the iteration number and see what happens.
 
As you mentioned in your comment your loop is big and nested, so there might be something else going on. Try to get rid of all extra operations and see what happens. 
 
Hope this helps, Cheers!