问题
When profiling my app with Instruments (looking for memory leaks), I get extremely different results with the iOS 5 iPhone Simulator from those I get with my iPhone 4 running iOS 5. The first picture shows the results from the profiling with the real device, and the second is with the simulator:
Real device:
iOS 5 Simulator:
This profile is taken up to the same point in the app in both cases: completion of viewDidLoad in the rootViewController's view lifecycle. I have waited in both of them for the total allocated memory to stable out. As you can see in the device graph, there are some extreme fluctuations occurring at about 00:10, which aren't present in the Simulator. On the real device, total allocated memory, at around 00:08, jumps from 1MB to 3.5 MB then back down to 1.5 MB and finally jumps to 4.74, where it stabilizes. The allocated memory for the Simulator is much more linear, with it climbing steady and quickly to around 2.35 MB where it stabilizes.
Another thing to note is the presence of 2.25 MB of allocated memory present on the device but not the Simulator from malloc and 700+ KB from CFNumber. As I am relatively new to using Instruments and profiling, I'm not exactly sure if this is normal. A quick Google search turned up nothing definitive. That 2.25 MB and 700 KB more than make up for the difference in memory allocation. To balance things, there are more entries for malloc with different amounts of memory present in the Simulator test not present in the device test.
Also, I found that when a second UIViewController is pushed onto the UINavigationController stack, allocated memory jumps to about 8.5-9 MB on the real device, but only about 4.5 to maybe 4.5 megabytes tops on the Simulator.
I know it is to be expected that the device would perform much differently from the Simulator, but should memory allocation not be pretty similar because the same code is being run on both devices? I would understand if this is a performance profiling, but for memory allocation, it seems that the numbers should be pretty similar. Can anyone shed some light as to whether this is normal or not?
回答1:
This behavior is to be expected. Technically, when you run profiling with the simulator you are measuring stats based on your desktop's hardware. Even if you're just profiling allocations you can't expect them to work similarly because a lot of software optimizations/algorithms/etc are based on the hardware it's running on.
Unfortunately, Apple doesn't have an iOS emulator. You're better of profiling with the device though, as emulators tend to still be unreliable and slow (e.g. Android emulator).
回答2:
You should always run leaks on an iOS device and never on the simulator. The results you get from the simulator will only serve as a distraction since they are rarely 100% accurate. You'll find yourself chasing down a lot of red herrings! hehehe
回答3:
I know it is to be expected that the device would perform much differently from the Simulator, but should memory allocation not be pretty similar because the same code is being run on both devices? I would understand if this is a performance profiling, but for memory allocation, it seems that the numbers should be pretty similar. Can anyone shed some light as to whether this is normal or not?
Technically the code is completely different. The simulator application compiles for x86 bytecode, while the device compiles for armv6/armv7.
来源：https://stackoverflow.com/questions/8468662/profiling-memory-leaks-with-instruments-huge-difference-between-iphone-4-and-io