Focus On: WebApp vs Native, Performance

With the recent additions of HTML5, web apps narrow the gap with their native counterpart on the functionalities debate.
An argument frequently used by the web apps partisans is that, as technology progresses and devices become more powerful, the gap
in performances will shortly no longer be an issue, bringing web apps to the level of native apps.
We noticed that, although as of right now, no one contests that currenty native apps offer better performance, there is no data to judge how important the difference is. This is what we evaluated here, by benchmarking CPU, graphics and I/O for web, native and hybrid (webviews) examples on iOS and Android.
In fact, for Android, hybrid and web apps offer the same performance since Just In Time compilation is activated on both, contrary to just web apps for iOS.
We also compared real life examples of apps with web and native versions.

Applidium is a Paris based mobile development agency - find us at http://applidium.com

Transcript

1. Paris, December 11, 2012 Focus On: WebApp vs Native, Performance Adrien Couque, Patrick Nollet
2. Focus On: WebApp vs Native, Performance 2012 • Mobile development is about UX Which technology should we use?
3. Focus On: WebApp vs Native, Performance 2012 • Web App HTML + Javascript Run inside a browser or a web view inside an application (hybrid approach) Few OS-specific UI elements Native Compiled language Run as a standalone program Direct access to the hardware
4. Focus On: WebApp vs Native, Performance 2012 •
5. Focus On: WebApp vs Native, Performance 2012 • Performance Bottlenecks •CPU •Graphics •I/O
6. Focus On: WebApp vs Native, Performance 2012 • CPU
7. 2012 • Compression Focus On: WebApp vs Native, Performance API zlib java.util.zip zlib.js 3 Mo pure text ≈ 3.000.000 characters 10 iterations (compression + decompression)
8. 2012 • Compression Focus On: WebApp vs Native, Performance x13 x23
9. 2012 • SHA-1 Focus On: WebApp vs Native, Performance API CommonCrypto MessageDigest crypto.js (Google)
10. 2012 • SHA-1 Focus On: WebApp vs Native, Performance x360 x180
11. 2012 • Pi Pure arithmetics (additions, multiplications...) Spigot algorithm (Gibbons) 100.000.000 iterations Yields ≈20.000.000 decimals Focus On: WebApp vs Native, Performance
12. 2012 • Pi Focus On: WebApp vs Native, Performance x16 x3
13. Focus On: WebApp vs Native, Performance 2012 • Graphics Vector Graphics UI Layout 3D
14. Focus On: WebApp vs Native, Performance 2012 • Vector Graphics API Core Graphics Canvas <canvas> Test: Draw a 7-degree Koch snowflake
15. 2012 • Vector Graphics Focus On: WebApp vs Native, Performance x10 x1.7
16. Focus On: WebApp vs Native, Performance 2012 • UI Layout API UICollectionView GridView CSS (float left) Test: display a grid of thumbnails images
17. 2012 • Layout Focus On: WebApp vs Native, Performance x10 x3
18. Focus On: WebApp vs Native, Performance 2012 • I/O Network Disk
19. Focus On: WebApp vs Native, Performance 2012 • SQL queries API sqlite 3 Web SQL Database (deprecated) Test: Count the occurrences of a word in a book using INSERT and SELECT statements
20. 2012 • SQL Focus On: WebApp vs Native, Performance x1.5
21. Focus On: WebApp vs Native, Performance 2012 • Real-life examples
22. Focus On: WebApp vs Native, Performance 2012 • Google Maps Native Frame Rate: Web Frame Rate: 59 fps 37 fps
23. Focus On: WebApp vs Native, Performance 2012 • Cut the Rope Native Frame Rate: Web Frame Rate: 10 fps 58 fps in-game menu loading screen
24. Paris, December 11, 2012 Applidium 17 rue du Faubourg du Temple 75010 Paris http://applidium.com Thank you