May I suggest to include
PassMark in benchmarking?
It appears to be a good suite of tests and gives detailed results.
CF-Bench could be interesting, too, as it compares native vs. Java.
By the way, I suspect that Vellamo Metal might be single-threaded, but I'm not sure. Anyway, single-threaded performance is important, too. I doubt that many Android apps are optimized for multicores.
Here are detailed Metal, PassMark and CF-Bench results from an Ainol Aurora 2 (Amlogic AML8726-MX):
Vellamo Metal 1.png (135.07K)
Number of downloads: 55
Vellamo Metal 2.png (134.1K)
Number of downloads: 52
Screenshot_2013-01-22-20-54-37.png (115.82K)
Number of downloads: 48
Screenshot_2013-01-22-20-54-47.png (104.63K)
Number of downloads: 44
Screenshot_2013-01-22-21-14-34.png (88.26K)
Number of downloads: 43
Screenshot_2013-01-22-21-14-41.png (92.75K)
Number of downloads: 40
The Hero 2 performs worse than this in all Metal tests save for RamJam.
Another interesting item is the total available memory. While most tablets have 1 GiB of RAM (with 2 GiB on the rise), not all of that is available to apps. (Note that this is not the currently free/available memory, but the total memory seen by the OS!)
Some benchmarks report the total memory, see the PassMark screenshot above for an example. However, there's a catch here - it's not 899 MiB (binary units of 1024) but 899 MB (decimal units of 1000).
Check the virtual file "
/proc/meminfo" if you're unsure, the unit there is "kB" (KiB, binary).
899 MB = 878 MiB.
So, in case of the Aurora 2, we have
878 MiB / 1024 MiB =
85.8% total memory available.
For the Hero 2, it's only
719 MiB / 1024 MiB =
70.2% total memory available.
I wonder what occupies those nearly 30% of RAM on the Hero 2 (GPU framebuffers?) and whether it's the same with the Venus (probably).
One more thing:
Screen refresh rate. On the Hero 2 it's apparently 62.95 Hz as shown by Vellamo's system info. On the Aurora 2 it's only 50 Hz, confirmed by Vellamo and the PassMark simple 3D test.
This post has been edited by Tzul: 22 January 2013 - 05:28 PM