• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Quadrant score goes up and up and up and up???

irvs101

Newbie
Hey all

So I killed all my processes and decided to run Quadrant and got 1146 and I thought okay maybe I shouldn't have killed the processes because they would be restarting while quadrant ran, so I restarted it again and what do you know I get 1212, so I left it 5 minutes and came back and run it again and get 1348. At this point I'm freaking bewildered because I've literally not changed anything and I run Quadrant again and what do you know I get 1436.

This leaves me with one of two opinions.

A) The Desire is such a 'smartphone' that it actually gets smarter as it gets older :)

B) Quadrant benchmark results should be taken with a pound of salt?

Oh and one more time before I go, just pulled a 1489, Not bad for a device that quadrant says should bench at 500

Opinions? Oh and what do you score on Quadrant?
 
Quadrant scores do get better. With some schedulers like on demand, the first few runs "open 'er up" but I even see this affect even on the performance governor.

They should be taken with a pinch of salt, definitely. Its no reflection on real world usage. The phones with the best scores could be slower than ones with much lower.

Part of the score is also based on Graphics too, so a good graphics chip will give a better score, but will see little real world merit in a home screen environment.

500 Is right for the Desire on 2.1, however its about 1100 for 2.2

On Defrost (basically the froyo equivelant of gingervillain) with certain custom kernels (Think it was a Vork at the time), I got something like 2055 in quadrant. When I had data2sd , I had a score of 2500, although this was fake I/O enhancement caused by cache write back.

Gingerbread custom roms are giving lower scores than their froyo equivelent, mainly due to the graphics not being optimised (although the phones feel snappier). I have just done a few tests and my best was 1564

I wouldnt put too much stock in it to be honest, although it can be useful to test different kernels on your own device but not useful to compare different phones.
 
So Quadrant is really a tool that tests your graphics processing capability?

Makes sense considering all the image tests it goes through.
 
Not only that. Quadrant tests CPU, data I/O, 2D & 3D graphics performance, etc. You get scores for different sections (5-6 of them?) and they all add together to make the total.

For example, SGS had a lower score than Desire, despite the fact that it had about 4-5 times better GPU.
 
No it tests CPU and I/O too. But all adds up to final score so doesn't necessarily mean anything
 
So what's the point of benchmarking software if it's not a tool for comparison with other devices? Isn't that the point?
 
That is indeed the point. Some good review sites use multiple benchmarks for example, because the benchmarks can be manipulated, and they sometimes are optimized only for certain manufacturer phones.
 
Back
Top Bottom