Well for all it's worth... I remember a promo pic for the Galaxy 2 at MWC: http://st.gsmarena.com/vv/reviewsimg...marena_009.jpg
Claiming a 5x performance increase over the SGX540 from their Quad core Mali-400 will obliterate the ULV Geforce since the tegra2's chip can barely beat the SGX540 as it stands now. Of course this isn't real life performance numbers, samsung is just probably simply quoting how many triangles/sec each setup can push and taking the ratio, which doesn't always translate to real world usage.
Either way, I have my money on samsung's quad core mali-400 over tegra 2. I have an Atrix, quite frankly it sucks, way below expectations. My Nexus S underclocked at 800mhz outperforms it in every 3D game I've tried. The Galaxy 2 will be my next phone. **** locked bootloaders and ****ty LCD screens, I'll live with slightly worse GPU performance for better ROM availability and SAMOLED plus. =)
I just hope the poor tegra 2 performance in games is due to the games not being optimized for Tegra right now. I've tried a few of the tegra optimized games as well and they don't seem particularly better than other games, though...
The score would be even higher at lower screen resolution.
It just proves that benchmarks can be unreliable and it's better to test performance during
normal usage of the released product.
GL benchmark Odroid-A tablet with Exynos 4210 (Exynos 4210 is used in SGS2):
GL benchmark Galaxy S:
Motorola Xoom and Atrix: