3 ms·

C code runs lightyears faster (aka. uses less power) than java. First of all, it's not lightyears, usually only a factor of two. Secondly, iOS apps are usually

C code runs lightyears faster (aka. uses less power) than java. First of all, it's not lightyears, usually only a factor of two. Secondly, iOS apps are usually written in Objective C, which uses dynamic binding, sacrificing part of C's performance. 5) iOS uses the GPU rather than the CPU to do most rendering tasks, So does Android: http://developer.android.com/guide/topics/graphics/hardware-accel.html#determining http://developer.android.com/guide/topics/graphics/hardware-... 8) Timer coalescing allows the CPU to sleep more The Linux kernel has timer coalescing (since 2.8.28).

lumpypua 12y ago
First of all, it's not lightyears, usually only a factor of two. That's true if you're running java on the jvm, android uses dalvik which is nowhere near as performant. https://code.google.com/p/android-benchmarks/ https://code.google.com/p/android-benchmarks/
- Yetanfou 12y ago
  Before quoting performance comparisons it is a good idea to check whether they are up to date. That page compares a very early version of Dalvik (1.6 a.k.a. 'Cupcake') to native C and a server JVM with a smart-ish jit compiler ('hotspot'). The first release of Dalvik which contained a jit compiler was 2.2 ('Froyo'), comparisons for that versions with Java have been performed by Sun/Oracle [1] showing it to be about half as fast as Java SE embedded 1.6.0_21. Both the JVM as well as Dalvik have made speed gains in the intervening 3.5 (!) years. As of Android 4.4 there is an experimental new runtime (prosaically named 'ART' for 'Android RunTime'...) which pre-compiles as much of the code as possible to native code. The speed gains currently are subject of much debate.
- fleitz 12y ago
  Haha no kidding, only a factor of two... thats not much faster... If you consider that it generally takes about 1 year for mobile processors speed to double I'd call that a lightyear.

lumpypua 12y ago

First of all, it's not lightyears, usually only a factor of two. That's true if you're running java on the jvm, android uses dalvik which is nowhere near as performant. https://code.google.com/p/android-benchmarks/ https://code.google.com/p/android-benchmarks/

Yetanfou 12y ago
Before quoting performance comparisons it is a good idea to check whether they are up to date. That page compares a very early version of Dalvik (1.6 a.k.a. 'Cupcake') to native C and a server JVM with a smart-ish jit compiler ('hotspot'). The first release of Dalvik which contained a jit compiler was 2.2 ('Froyo'), comparisons for that versions with Java have been performed by Sun/Oracle [1] showing it to be about half as fast as Java SE embedded 1.6.0_21. Both the JVM as well as Dalvik have made speed gains in the intervening 3.5 (!) years. As of Android 4.4 there is an experimental new runtime (prosaically named 'ART' for 'Android RunTime'...) which pre-compiles as much of the code as possible to native code. The speed gains currently are subject of much debate.
fleitz 12y ago
Haha no kidding, only a factor of two... thats not much faster... If you consider that it generally takes about 1 year for mobile processors speed to double I'd call that a lightyear.