GPU Performance? Lack Luster?

[Zzim]

Is anyone else noticing that the GPU in this phone is very slow when comparing it to some other Galaxy S and Galaxy S II models? Previously I had a Samsung Epic 4G and the GPU in smart bench 2011 would score ~2200. Now the Epic 4G has the same PowerVR SGX540 as the galaxy nexus so it should score and perform just as well.

Now I also have a Galaxy S II Epic Touch 4G and the specs for that phone are similar to the Galaxy Nexus. The Epic Touch has a Mali 400 and scores ~2500. Now the the Epic Touch will play N64oid just fine with no lag at all where the Epic would lag a bit.

When playing N64oid on the Galaxy Nexus its just a slow and laggy as my old droid charge was. Now the Galaxy Nexus will not score above ~1800 in smartbench. The droid charge scores higher than that.

Any ideas can any one post their scores so I can see if its just my phone?

 

[silow]

Keep in mind the GN's GPU is rendering over 2 times more pixels.

Disclaimer: I don't know if Smartbench runs their benchmarking at native resolutions.

 

[Zzim]

Good point....

Can any one run smart bench 2011 and post a screen shot so i know its not my device?

 

[theboo7]

Sent from my Galaxy Nexus using XDA App

 

[Zzim]

Well I guess its not the phone......damn.....

 

[th0r615]

Don't put your faith in stretchmarks. At the end of the day if one GPU scores higher then the other on smart bench but both run the games you wanna play with no problems then whats the point of having the bigger number?

 

[@rbiter]

And are you running @ 308 or 384?

Sent from my Galaxy Nexus using xda premium

 

[zephiK]

Don't put your faith in stretchmarks. At the end of the day if one GPU scores higher then the other on smart bench but both run the games you wanna play with no problems then whats the point of having the bigger number?

Same reason why people overclock their computers / obsessed about not getting 100+ fps. Oh the eye can't notice any difference above 30 fps but people want 250 fps anyway.

Keep in mind that the benchmark that you're using is not optimized for ICS. So its a unaccurate result. Benchmarks don't mean anything. It's real-time performance that really matters.

 

[guthrien]

Oh the eye can't notice any difference above 30 fps but people want 250 fps anyway.

Most people seem to notice the difference between say, 30 and 60fps in video. The eye is pretty capable of noticing sluggishness because it happens in contrasts, even if they can't count each frame.

 

[bash_array]

Most people seem to notice the difference between say, 30 and 60fps in video. The eye is pretty capable of noticing sluggishness because it happens in contrasts, even if they can't count each frame.

Lol. The human eye cannot discern above 30 fps. 31fps and 60fps look the same.

Sent from my Galaxy Nexus using xda premium

 

[bash_array]

Correct me if I'm wrong but the gnex is software rendered not hardware so it will bench out badly. There are roms out now that force hardware rendering... the quadrant scores jump from 1500 to 2800

Sent from my Galaxy Nexus using xda premium

 

[guthrien]

Lol. The human eye cannot discern above 30 fps. 31fps and 60fps look the same.

Sent from my Galaxy Nexus using xda premium

Lol (?) I can't seem to find the scientific or even opinionated "proof" that people can't tell the difference. Even subjective experience tells me that people can tell the difference in displays so I don't know why it's treated as a glib fact by you. There seems to be a lot of tests and experiments out there just by doing a cursory search. Can you point me to the definitive evidence?

 

[xNEM3S1Sx]

Lol. The human eye cannot discern above 30 fps. 31fps and 60fps look the same.

Sent from my Galaxy Nexus using xda premium

If the image switches between black and white each frame, the image appears to flicker at frame rates slower than 30 FPS (interlaced). In other words, the flicker fusion point, where the eyes see gray instead of flickering tends to be around 60 FPS (inconsistent). However, fast moving objects may require higher frame rates to avoid judder (non-smooth, linear motion) artifacts — and the retinal fusion point can vary in different people, as in different lighting conditions. - Wikipedia

For object motion, 30 vs 60 can be a huge difference, most noticeable when turning.

 

[Zzim]

Don't put your faith in stretchmarks. At the end of the day if one GPU scores higher then the other on smart bench but both run the games you wanna play with no problems then whats the point of having the bigger number?

I guess that's my issue N64oid lags a lot and does not have smooth game play when in goran mountain in major's mask or just in general, but on my Galaxy S II Epic Touch 4G I get no lag at all.....none...sucks that sprints 4g and 3g blow

 

[Zzim]

Correct me if I'm wrong but the gnex is software rendered not hardware so it will bench out badly. There are roms out now that force hardware rendering... the quadrant scores jump from 1500 to 2800

Sent from my Galaxy Nexus using xda premium

Sooo would this be the same for any app or game? Its going to render using the CPU(software) and not use the GPU....? That sounds a bit strange but in this case would be the reason why it runs so slow.