You also talk about different operating systems being different.
So you want to compare x86 Windows to ARM Android? Or what?
x86 Windows vs ARM Android is a far more accurate comparison when looking at power usage which is what this article claims to be. Windows 8 versus Windows RT shows us nothing, I could make a ROM for Android that looked like Windows 8, on a binary level it would probably be about as close as x86 Windows vs Windows RT. Ignoring that though, run Android on both. Intel has their own builds of Android. The basis of this article is performance versus power usage, Anand makes shockingly ignorant claim that Intel is looking good in that aspect based on a laughably terrible comparison.
OS choice: Covered by availability. THERE ARE NO OTHER COMPARABLE OPERATING SYSTEMS.
They didn't use a comparable operating system, not even close. They chose a crippled OS for ARM versus the driving OS for all of x86. If you want to look at power usage models, which is the obvious lie they present this article under, then compare it to an OS that is capable of using the power saving features of the horribly out of date SoC you chose to rig the test with.
SoC choice: What Intel sent (and what is in use). With FUTURE WORK TO BE DONE ON OTHER SOCs.
Covered.
No, it isn't. If Anand handled his launch review for the 680 GTX against the 6970 this forum would have been in a fit of rage. That would have been closer to even then what Anand did in this test. The S4 Pro has been out longer then the 7970 was when 680 GTX hit. I don't think you comprehend how crooked this test was and how badly rigged it was to prove their agenda.
Lack of GPU performance numbers:
"While we don't yet have final GPU benchmarks under Windows RT/8 that we can share numbers from..."
"NVIDIA's GPU power consumption is more than double the PowerVR SGX 545's here, while its performance advantage isn't anywhere near double."
Covered.
The GPU power numbers are less then meaningless without performance numbers. That isn't remotely in the league of being covered. Again, they should be going against the S4 Pro, not the Tegra 3, for this comparison anyway. Rigged test is rigged.
He used an SoC which is in 75% of shipping (as far as I can tell) Windows RT devices.
Based on early sales estimates Windows RT for the *year* is going to hit about one quarter of what Android will sell *today*. This is supposed to be about ARM vs x86- I don't understand why we are discussing a miniscule niche OS as being the driving force behind comparing hardware architectures.
WinRT is also probably optimised for Tegra 3
The power saving core *DOES NOT WORK*. A main driving force behind the entire SoC is non functional, and you are going to try, with a straight face, to say that is optimized? This article was supposed to be comparing x86 to ARM, Tegra 3 shouldn't have been used at all, but then they take that seriously outdated SoC and cripple the hell out of it for their comparison? Obnoxiously stupid PR stunt, it is not defensible.
Tegra 4 is also indicated to have a companion core, if that works in the same way as Tegra 3, it will also not work under Windows RT.
BlackBerry is more important then Windows RT. People can try and rub one out over MS's tablet OS, it has been a profound failure. Right now Android and iOS dominate the market, why are we looking at benches from a failed OS?
If Anand reviews the 780 and 8970 under Linux are you going to try and defend that too? You may think that is extreme, and you would be right, because Linux has a higher share of the desktop gaming market then WinRT has of the tablet market.
Covered doesn't mean that everything has been fully investigated to find the truth, it means it's been discussed and either explained or noted as an issue, or it's just blindingly obvious.
Covered would be saying *I'm working as a paid shil for Intel*. Other then that it wasn't covered, it was just an attempt at explaining why they are wasting bandwidth with useless information.
Facebook
Twitter