View Single Post
Posts: 840 | Thanked: 823 times | Joined on Nov 2009
#341
Originally Posted by Kangal View Post
Ohkay.

I didn't use any of those weblinks....I ran the benchmark off my phone.
And I'm more inclined to believe my hands-on score of ~1,500 rather than one published by an unknown. And like I said, I can get a better score if I tried.
well I'm certainly not making this **** up. If that's what you think

http://www.flickr.com/photos/umpcportal/6329289344/
All phones are tested stock, trying to get high scores by tweaks isn't the norm for any of these benchmarks. These are out the box devices.

My point still stands, benchmarks mean nothing when it comes to real performance. ibrakalifa is possibly true that Harmattan running on a more modern processor would outpace even these high-end devices.

(The only thing benchmarks can do is to prove one device is actually faster than another, when the scores are very far apart.
I'm talking like a difference between gingerbread SGS and SGS2 where they behave close/slight edge to S2...but in benchmarks the S2 proves its superiority by showing more impressive figures.)
I think benchmarks were discussed earlier in this thread so we are covering already covered ground. Usually I would agree for a benchmark that uses hardware not currently supported by any software other than the benchmark itself but again sunspider is a realworld test which explains why ibrakalifa might be showing a better score with lower hardware. Which was what his assertion was based on anyway, his sunspider score. Sunspider is a realworld test that takes these things into account ( eg OS and browser) and is based on actual javascript tasks you would encounter in actual sites. It's not just an arbitrary figure based on theoretical potential of the hardware. It's a timed actual task on a site. Did you time the "slight edge" of the S2? to know if this slight edge matchs the benchmark difference? It seems you accept a qualitative difference but dispute the quantitative test of it.