Xcode Hardware Performance
These are the results from running Xcode on a non-trivial open source project using various Macs. The goal is to give developers a relative idea of how each computer model compares to one another. Read the specifics and contributing sections for more info.
Xcode 9
| 🖥 | Computer Model | CPU | RAM | Fresh Build Time | Incremental Build Time | Date | New Build System? |
|---|---|---|---|---|---|---|---|
| 💻 | iMac Pro 10 Core, Late 2017 |
3 GHz Intel Xeon W | 64 GB | 0:41 | 0:07 | 12/28/17 | ❌ |
| 💻 | MacBook Pro Retina, 15", 2017 |
2.9 GHz i7 | 16 GB | 0:49 | 0:16 | 10/9/17 | ✔️ |
| 💻 | MacBook Pro Retina, 15", 2017 |
2.9 GHz i7 | 16 GB | 0:50 | 0:15 | 10/9/17 | ❌ |
 |
Mac Mini Mid 2012, 512 SSD |
2.3GHz Quad-Core i7 | 16GB | 1:32 | 0:18 | 10/20/17 | ❌ |
| 💻 | MacBook Pro Retina, 15", 2017 |
2.8 GHz i7 | 16 GB | 1:50 | 0:14 | ❌ | |
| 💻 | MacBook Retina, 15", Mid 2012 |
2.6 GHz i7 | 8 GB | 2:26 | 0:23 | ❌ | |
| 💻 | MacBook Pro 13", Mid 2012 |
2.9 GHz i7 | 8 GB | 2:30 | 0:23 | ❌ |
Xcode 8
| 🖥 | Computer Model | CPU | RAM | Fresh Build Time | Incremental Build Time | Date |
|---|---|---|---|---|---|---|
| 🖥 | iMac 5K 512GB Flash, 27", Late 2015 |
4.0Ghz i7 | 32 GB | 0:28 | 0:06 | |
 |
Mac Pro Flash Storage, Late 2013 | 3.5 GHz 6-Core Xeon E5 | 32 GB | 0:30 | 0:05 | |
| 💻 | MacBook Pro Retina, 15", Mid 2015 |
2.8 GHz i7 | 16 GB | 0:39 | 0:07 | |
| 💻 | MacBook Pro Retina, 15", Mid 2015 |
2.5 GHz i7 | 16 GB | 0:42 | 0:09 | |
| 💻 | MacBook Pro Retina, 15", Mid 2012 |
2.3 GHz i7 | 16 GB | 0:46 | 0:08 | |
| 💻 | MacBook Pro Retina, 15", Early 2013 |
2.4 GHz i7 | 8 GB | 0:47 | 0:10 | |
|
Mac Mini Mid 2012, 512 SSD |
2.3GHz Quad-Core i7 | 16GB | 0:50 | 0:09 | |
| 💻 | MacBook Pro Retina, 15", Mid 2014 |
2.2 GHz i7 | 16 GB | 0:51 | 0:07 | |
| 💻 | MacBook Pro Retina, 15", Late 2013 |
2.3 GHz i7 | 16 GB | 0:53 | 0:09 | |
| 🖥 | iMac HDD 1 TB, 27", Late 2012 |
3.4 GHz i7 | 8 GB | 0:54 | 0:12 | |
| 💻 | MacBook Pro Touch Bar, 15", Late 2016 |
2.9 GHz i7 | 16 GB | 0:54 | 0:17 | |
| 💻 | MacBook Pro 15", Early 2011 |
2.0 GHz i7 | 8 GB | 1:00 | 0:13 | |
| 💻 | MacBook Pro 15" Early 2011, SSD |
2.0 GHz i7 | 8 GB | 1:01 | 0:10 | |
| 💻 | MacBook Pro Touch Bar, 15", Late 2016 |
2.9 GHz i7 | 16 GB | 1:04 | 0:26 | |
| 💻 | MacBook Pro 13", Early 2015, 1TB SSD |
3.1 Ghz i7 | 16 GB | 1:25 | 0:09 | |
| 💻 | MacBook Pro Retina, 13", Early 2015 |
2.7 GHz i5 | 8 GB | 1:35 | 0:11 | |
| 💻 | MacBook Retina, 12", Mid 2017 |
1.4 GHz i7 | 16 GB | 1:49 | 0:16 | |
| 💻 | MacBook Pro 13”, Mid 2012 256 SSD |
2.5 GHz i5 | 8 GB | 2:08 | 0:19 | |
| 💻 | MacBook Air 13", Mid 2012, 512 SSD |
2 Ghz i7 | 8 GB | 2:10 | 0:11 | |
| 💻 | MacBook Retina, 12", Early 2016 |
1.2 GHz m5 | 8 GB | 2:10 | 0:12 | |
| 💻 | MacBook Pro 13”, Late 2013 256 SSD |
2.4 GHz i5 | 8 GB | 2:17 | 0:23 | |
| 💻 | MacBook Retina, 12", Early 2015 |
1.2 GHz M | 8 GB | 2:28 | 0:12 | |
| 💻 | MacBook Retina, 12", Early 2015 |
1.1 GHz M | 8 GB | 3:00 | 0:12 | |
| 💻 | MacBook Pro 13", Early 2011, 512 SSD |
2.7 Ghz i7 | 8 GB | 2:30 | 0:16 | |
| 💻 | MacBook Air 13", Mid 2013, 128 SSD |
1.3 Ghz i5 | 4 GB | 2:58 | 0:24 | |
| 💻 | MacBook Pro 13”, Mid 2010 512 SSD |
2.4 GHz Intel Core 2 Duo | 8 GB | 3:01 | 0:24 | |
| 🖥 | iMac HDD 500 GB, 21.5", Mid 2010 |
3.06 GHz i3 | 12 GB | 3:59 | 0:14 | |
| 🖥 | iMac HDD 1 TB, 21.5", Late 2012 |
2.7 GHz i5 | 8 GB | 4:15 | 0:33 |
Specifications
For the test, I decided to use an app that I actually work on: eidolon.
For "fresh" builds, I cleaned the build folder (⌘⇧K) repeatedly until it worked with no permissions problems. Then I sat and waited for Xcode to index the project. I also made sure the simulator was closed, so these times include booting the simulator and launching the app. Then I hit ⌘R and start a timer, only ending it when the app had fully launched.
"Incremental" builds represent a more common use case: changing one file and recompiling with the simulator already running. I added print("hello!") to application(: didFinishLaunchingWithOptions:) and hit ⌘R, timing the time it took for the app to launch.
I repeated each test a few times and took their average times.
Contributing
It would be super-cool if we could perform the above tests on a variety of machines and consolidate the results here. You can follow the instructions to download the code and the project dependencies, and send a pull request adding your own results. I'd super-appreciate it! 🙇
Please note that this project is released with a Contributor Code of Conduct. By participating in this project, you agree to abide by [its terms](/Code of Conduct.md).
License
This work is licensed under a Creative Commons Attribution 4.0 International License.