Benchmarks and other comparisons/impressions

[Shocky]

Translation - the graphics benchmark in Smartbench is a stress test.

All graphics benchmarks are designed to be stress tests, otherwise what would be the point running them?

That doesnt explain why a GPU that's obviously faster scores less in smartbench.

 

[EarlyMon]

The inference is clear, the gpu is only obviously faster for some uses of the graphics API. Depending on the specific stress, it's not faster.

Neither the favorable nor unfavorable benchmarks matter for your use beyond what the app devs for games you prefer will use. Those devs would do well to use the benchmark results in choosing their calls wisely. Benchmarks are tools for devs, not horse races - the results must always correlate to use.

 

[Shocky]

The inference is clear, the gpu is only obviously faster for some uses of the graphics API. Depending on the specific stress, it's not faster.

Obviously? You don't know that's even the case here, that's an assumption based on results from one benchmark.

 

[EarlyMon]

All I did was quote you without sarcasm or hostility on the phrase obviously faster.

No need to read into my words, honestly.

I only have objected to uncorrelated benchmarks getting in the consumer's way of making good choices. That's it, my whole agenda. If yours is different, I can't be of help to you.

 

[mrqs]

Obviously? You don't know that's even the case here, that's an assumption based on results from one benchmark.

do you have another explanation for those results?

 

[Shocky]

do you have another explanation for those results?

Yes, there

 

[mrqs]

Yes, there

 

[nateap87]

the way i see it, the sensation is more than sufficient enough for what ANYONE will be using it for. the sgsII is like having a car that can go 200mph, but chances are, you'll never utilize it. there aren't any games or apps that would run better on one or the other. they'll run to their full potential on the sensation.

 

[Shocky]

New version of Nenamark2 is out

Galaxy S II is ahead by a huge margin, I scored 49.1 fps

Sensation scores 26.06 fps but is at a disadvantage due the resolution.

Nena Innovation AB - NenaMark results

other than it using a lot of certain api calls that happen to be faster on the apparently generally slower chip

Or it happens to be faster on other hardware or slower on others when coded in a certain way, which is more likely.

I'm not really seing much point in the game test in smartbench right now, waste of time running it.

 

[EarlyMon]

Seems pretty comprehensive and trustworthy - I got a whopping 4.9 fps on my Evo!

Yeah - fast graphics it never had.

Trusting Nenamark2, I'd expect that the Sensation might degrade by a 1/3 at best due to resolution differences - maybe get around 33 fps all things being equal. So, either Sense or the 8x60 is pulling it below the SGS2, imo only.

 

[Shocky]

Recent firmware upgrades have improved 3D performance on the Galaxy S II, Electopia results have increased from 37 fps to 50 fps and GLBenchmark went from 41 fps to 59 fps. (limited only by vsync here)

Oddly enough smartbench shows no improvements, big supprise there.

 

[rich2626]

Old thread I know...

but I ran two tests on my Sensation....on NenaMark 2,I got 45.5fps,and on quadrant I got 2700,it seems rather high for a old phone..are these tests actually accurate,or just really a novelty factor?