Kishonti GFXbench

[Ailuros]

By the way can someone with leads to Kishonti see if there are any mistakes for the following cases:

http://www.glbenchmark.com/phonedetails.jsp?D=Alcatel+One+Touch+995&benchmark=glpro25

An Adreno"205" clocked at what?

http://www.glbenchmark.com/phonedetails.jsp?D=HTC+Thunderbolt+%28Mecha%29&benchmark=glpro25

Another Adreno205 wonder-puppy...

Or even better those two pose even above Adreno320s: http://www.glbenchmark.com/result.j...ersion=all&certified_only=1&brand=all&gpu=all

Also that one: http://www.glbenchmark.com/phonedetails.jsp?D=Rikomagic+MK802+II&benchmark=glpro25 has quite interesting 1080p results compared to all other tests. Given the onscreen 2.5 results and 2.1 off-/and on-screen results I'm not so sure whether it actually yields 27.6fps in 1080p 2.5 or rather 2.76fps in reality.

--------------------------------------------------------------------------------------------------------------

On another note the first "octa-core" Vivante GC4000 scores are out: http://www.glbenchmark.com/phonedetails.jsp?D=Huawei+U9510+(GC4000)&benchmark=glpro25

 

[kishonti]

Thank you, Ailuros!
Removed them.

 

[Ailuros]

Thank you, Ailuros!
Removed them.

One is glad to be of service :smile:

------------------------------------------------------------------------------------------------------------------

Broadcom Videocore IV not bad at all:

http://www.glbenchmark.com/phonedetails.jsp?D=Samsung+GT-I9105P&benchmark=glpro25

 

[Ailuros]

We have another winner:

http://www.glbenchmark.com/phonedetails.jsp?D=Motorola+Droid+X2&benchmark=glpro25

A Tegra2 that yields over 520 fps in Egypt 2.5 offscreen

 

[Simon F]

A Tegra2 that yields over 520 fps in Egypt 2.5 offscreen

If a pixel falls in the woods and no-one sees it ... does it need to fall?

 

[Mariner]

We have another winner:

http://www.glbenchmark.com/phonedetails.jsp?D=Motorola+Droid+X2&benchmark=glpro25

A Tegra2 that yields over 520 fps in Egypt 2.5 offscreen

It's probably attached to a heatsink the size of a house which is cooled by LN2.

 

[Rys]

It'd need about 125GHz (quick estimate!). Even with your big heatsink and lots of LN2, that's a big ask for 40LPG!

 

[rpg.314]

If a pixel falls in the woods and no-one sees it ... does it need to fall?

If people are recording the sound of the pixel falling...

 

[Ailuros]

Obviously no results anywhere but there's a "PowerVR RGX600" hidden somewhere in the Kishonti database

 

[Lazy8s]

Lots of interesting entries and results pop up from time to time. For instance, the Vivante GC1000 powered PXA988 Armada product from Marvell now tops the list in fill rate offscreen.

 

[Ailuros]

Thank you, Ailuros!
Removed them.

I know it's hairsplitting but is it possible to assign the iPhone4S to the SGX543 GPU category? http://www.glbenchmark.com/result.j...ce=1&version=all&test=448&gpu=PowerVR SGX 535

As it stands it looks like the SGX535 is damn fast when you select results by GPU. I don't want to calculate what kind of frequency it would need to reach 1326 frames

 

[french toast]

I know it's hairsplitting but is it possible to assign the iPhone4S to the SGX543 GPU category? http://www.glbenchmark.com/result.j...ce=1&version=all&test=448&gpu=PowerVR SGX 535

As it stands it looks like the SGX535 is damn fast when you select results by GPU. I don't want to calculate what kind of frequency it would need to reach 1326 frames

Ailuros..whats all this business about frames? Why are some greater than others? I thought the frames would be a constant and only the speed in which all frames in a demo are rendered by each gpu would change?

Confused :/

Edit. perhaps some gpus are unable to render every frame correctly and so jump on the next one..there for gettimg to the end of the demo..but having skipped a number of frames to do so??

 

[Rys]

The benchmark has a fixed runtime and the idea is you render as many frames as you can in that time.

 

[french toast]

The benchmark has a fixed runtime and the idea is you render as many frames as you can in that time.

That explains it then cheers

 

[Ailuros]

Edit. perhaps some gpus are unable to render every frame correctly and so jump on the next one..there for gettimg to the end of the demo..but having skipped a number of frames to do so??

Sounds way too complicated and rather dumb to me for a benchmark.

If you take the frame values of each result and divide it by the frames per second fps you get at all times an almost 113 divider. Tegra3 = 13.2 fps * 113 = 1491.6 frames, iPad4 = 51.9 fps * 113 = 5864.7 frames as examples. No idea why they're using it, since it would be times easier to use just the fps values for comparisons, but it sounds more like frames are a measurement for how many frames each GPU can process within the timeframe of 113 seconds.

 

[french toast]

Sounds way too complicated and rather dumb to me for a benchmark.

If you take the frame values of each result and divide it by the frames per second fps you get at all times an almost 113 divider. Tegra3 = 13.2 fps * 113 = 1491.6 frames, iPad4 = 51.9 fps * 113 = 5864.7 frames as examples. No idea why they're using it, since it would be times easier to use just the fps values for comparisons, but it sounds more like frames are a measurement for how many frames each GPU can process within the timeframe of 113 seconds.

Right so benchmark lasts for 113 seconds...not all gpus can complete all frames of benchmark..hence the lower frames score..cheers

 

[Rys]

It's not a matter of completing all frames. There is no fixed number of frames submitted to the GPU to render, it's dynamic based on how fast you can render them.

 

[Ailuros]

Right so benchmark lasts for 113 seconds...not all gpus can complete all frames of benchmark..hence the lower frames score..cheers

That's not what I said or meant. See also Rys' answer above.

 

[french toast]

but it sounds more like frames are a measurement for how many frames each GPU can process within the timeframe of 113 seconds.

mmm, ok.

Well I think I already get it...there is a 113 sec time limit and which ever gpu can complete as many frames as it can in that window gets the higher frames score... (which is then divided by the demo time in seconds to get the fps number)
If thats what you are both saying then I got it right the first time.. (except no fixed amount of frames..just time limit)

 

[Ailuros]

I give up.