Matthew Morey

I'm an engineer, developer, author, hacker, creator, tinkerer, traveler, snowboarder, surfer, and husband.

I create iOS apps professionally and independently.

Tools for Core Data

24 November 2013

This is the second in a series of articles on Core Data performance. Material is based on the talk High Performance Core Data I gave at CocoaConf in Atlanta on November 15th, 2013. An hour talk is not enough time to discuss such an advanced topic so I will be continuing the conversation here.

There is not one correct way to use Core Data as data models vary a tremendous amount from app to app. You should always start with the most basic Core Data stack, which includes a single managed object context on the main queue. Don't deviate from this stack until you encounter a performance problem. Usually you will know you have a problem because the UI will block or an operation, like an import, will take a long time to complete.

Core Data performance is a balance between memory and speed

Instead of taking a subjective guess on performance bottlenecks you really should measure and get exact results. Ideally you would measure, make specific changes, and measure again to validate the changes. Luckily we have tools to help us make detailed measurements.

Core Data performance is a balance between memory and speed


Tools don't have to be complicated. A simple pair of macros can tell you a lot about the performance of your app. A time delta is easily calculated by using timeIntervalSinceNow method on an NSDate object.

#define TICK NSDate *startTime = [NSDate date]
#define TOCK NSLog(@"Elapsed Time: %f", -[startTime timeIntervalSinceNow])

[self methodYouWantToMeasure];

Launch Arguments

Useful debugging output and behavior can be enabled by passing launch arguments into a target's scheme. By passing a launch argument you are overriding the current value in NSUserDefaults for that execution cycle. If using SQLite as the persistent store you have many options:

  • [1,2,3]: Allows you to see the actual SQL statements sent to the SQLite persistent store used by Core Data. Elapsed time for each query is also outputted. The verbosity of the output can be controlled by setting its value between 1 and 3. The higher the value the more SQL output.
SQLDebug launch argument
  • 1: Turns on SQL syntax coloring.

  • 1: Logs exceptions during data migrations.

  • [0,1,2]: Allows you to control how disk syncing is performed. Disk syncing is disabled when the value is 0. For normal behavior, set the value to 1. The default option is 2, which performs disk syncing via the fctl FULL_FSYNC command, which guarantees data is properly written to disk.

  • 1: Enables extra integrity checking.

  • [1,2,3]: Enables assertions for Core Data's multi-threading policy. Increasing values enables more restrictions.


You can view the actual SQLite database used by Core Data by browsing the app's Documents directory on disk. On the simulator the .sqlite file is stored deep within a users Application Support directory:

~/Library/Application Support/iPhone Simulator/7.0.3/Applications/XXXX-XXXX-XXXX-XXXX/Documents

You should switch out the 7.0.3 for whichever version of the simulator you are using. The XXXX-XXXX-XXXX-XXXX, which uniquely identifies your app, is automatically generated by Xcode.

In SQLite the database schema is stored in a special table named sqlite_master.

$ sqlite3 UFO.sqlite

sqlite> select * from sqlite_master;


The Z prefix is Apple's convention and should not be altered. Really you should never manually write to a SQLite Core Data database. Apple is free to change its underlining convention at anytime. If you find the need to manually manipulate the SQLite database behind Core Data's back you might want to forgo Core Data for raw SQL. There are a couple of popular frameworks that can help you manage raw SQL implementations, such as FMDB and FCModel.

For analyzing your data, there is nothing wrong with poking around the SQLite database file, just don't modify the contents. For example, data summation via the COUNT function is useful for determine how diverse your attributes are.



SQLite Professional

If you prefer a GUI interface for SQLite I recommend SQLite Professional. In addition to inline editing and normal queries you can also import and export with MySQL, JSON, or CSV. Its import functionality can be used to generate seed files, which can help avoid lengthy network calls on first launches of apps.

SQLite Professional tool SQLite Professional tool

Pony Debugger

Square's PonyDebugger offers an assortment of debugging tools. The Core Data browser feature allows you to registers your app's NSManagedObjectContexts and browse all entities in your data model. Although the data is read-only it is still a useful tool.

Pony Debugger

Installing PonyDebugger involves setting up a local server and integrating it with the desired app.

Core Data Editor

Core Data Editor allows you to view, edit, and analyze your Core Data backed Mac or iOS app. It supports XML, binary, and SQLite persistent store types. In addition to basic attribute editing it also allows you to edit relationships. Because it is familiar with Apple's naming conventions it removes the Z prefixes from the underlying store.

Core Data Editor

Debug Gauges

New with Xcode 5, Debug Gauges provide a quick way to view an app's resource consumption. The memory gauge is useful for Core Data as it is an indicator of how much of your object graph is living in memory.

When using gauges on the simulator for iOS apps the usage percentage is not helpful as it is calculated based on the available resource of the Mac system, not an actual device. You should only look at the total usage amount.

For iOS 7 only apps the available memory ceiling at the system level is 512 MB (iPhone 4, iPod touch 5th generation, and iPad 2). You should try to stay well below this amount.


Instruments is the tool of choice for investigating almost all Mac and iOS performance issues including Core Data problems.

Instruments Core Data template

The default Core Data template includes:

  • Core Data Fetches Instrument: Captures information, such as fetch count and duration, about fetch operations.

  • Core Data Cache Misses Instrument: - Captures information about fault events that result in cache misses.

  • Core Data Saves Instrument: - Captures information on save events.

You should also make use of the Faults instrument.

  • Core Data Faults Instrument: - Captures information on fault events that happen during lazy initialization of NSManagedObjects or relationships.

Here is an example of using the Core Data template to debug a NSFetchRequest that is causing app launch to take an unacceptable amount of time.

In addition to the Core Data tools the File Activity and Timer Profile tools provide valuable information when debugging Core Data.

When dealing with Core Data you should always use tools to take an objective measurement of performance. If SQLite is used as the persistent store you have several debugging options. In most situations the built in Core Data Instruments template is the best place to start.

Measure first, make specific and targeted changes, and measure again for validation.

If Core Data interest you please considered subscribing to my infrequent email newsletter about Core Data.