29k points in 3DMark is nearly 50% faster than what Mullins gets in the Discovery tablet without the powerplug.
I think it depends on what score was used for the Tegra
Clearly Xiaomi used the "Overall" 3DMark Ice Storm Unlimited score for A7 and Tegra K1. The pure graphics score would be higher in comparison.
The GPU performance difference in GFXBench T-Rex HD Offscreen between Mullins in a 11.6" tablet on AC power (where CPU/GPU frequencies are artificially boosted) and Tegra K1 in a thin 7.9" tablet is more than 60% in favor of TK1!
hmmm, what does that mean?
See techspot review data measurements page.
Clearly Xiaomi used the "Overall" 3DMark Ice Storm Unlimited score for A7 and Tegra K1. The pure graphics score would be higher in comparison.
The GPU performance difference in GFXBench T-Rex HD Offscreen between Mullins in a 11.6" tablet on AC power (where CPU/GPU frequencies are artificially boosted) and Tegra K1 in a thin 7.9" tablet is more than 60% in favor of TK1!
It’s worth noting that all the benchmarks were ran with the AMD Discovery tablet plugged in to AC power. When the tablet is benchmarked on battery power, I noticed a roughly 10% performance drop as the Mullins APU switches to using a more aggressive boost algorithm in attempt to lengthen battery life.
In comparison, you don’t see any performance drop on battery power with Intel’s Bay Trail Z3740. Obviously it’s not great to see a performance reduction on battery, but if it does lengthen battery life I might be kind to the way AMD deals with power gating. As I only had limited time with the Discovery reference tablet and couldn’t examine battery life on Mullins devices, I’ll wait until I get longer hands-on time with an OEM’s Mullins-based tablet before making any further judgements.
well Tegra runs also very well on Win RT...The main problem with comparing Tegra and Mullins will be that Tegra will be run only on Android and Mullins is only run on Windows, so it is hard to compare them.
Clearly Xiaomi used the "Overall" 3DMark Ice Storm Unlimited score for A7 and Tegra K1. The pure graphics score would be higher in comparison.
The GPU performance difference in GFXBench T-Rex HD Offscreen between Mullins in a 11.6" tablet on AC power (where CPU/GPU frequencies are artificially boosted) and Tegra K1 in a thin 7.9" tablet is more than 60% in favor of TK1!
well Tegra runs also very well on Win RT...
TK1 is shaping up to look like the Tegra Nvidia has always wanted.
I will wait for an official announcement before putting away Tegra off the next surface. TK1 is by far the best Nvidia SoC, with solid Windows drivers, and over 2 decades of good cooperation with Microsoft. I don't see any reason to discard them (of course it may be possible, but at this stage, nobody knows)True, but their still not really the same (from people I've talked to with WinRT tablets at least). Also might be a moot point if K1 doesn't end up in a version of surface anyway.
I don't read Chinese so, to be honest, nothing is clear about the numbers obtained, testing procedures, etc.
So 50% higher 3dMark seemed like a bit of an exaggeration.
The fact that Tegra K1's score comes from an actual soon to ship 7.9'' tablet, not a bulky 11.6'' Reference Platform ''tablet'' @ AC power makes it even more impressive. I doubt that this top-bin Mullins chip will find its way to smaller tablets, unlike Bay Trail-T and Tegra 4 (and probably Tegra K1/Cherry Trail-T too).
The fact that Tegra K1's score comes from an actual soon to ship 7.9'' tablet, not a bulky 11.6'' Reference Platform ''tablet'' @ AC power makes it even more impressive. I doubt that this top-bin Mullins chip will find its way to smaller tablets, unlike Bay Trail-T and Tegra 4 (and probably Tegra K1/Cherry Trail-T too).
The GPU performance difference in GFXBench T-Rex HD Offscreen between Mullins in a 11.6" tablet on AC power (where CPU/GPU frequencies are artificially boosted) and Tegra K1 in a thin 7.9" tablet is more than 60% in favor of TK1!
Then you pick a Baytrail tablet for Windows or the most reasonably priced tablet from a decent manufacturer for Android.
Wow, apparently reading comprehension is not always a strong suit for everyone Didn't I already tell you to read the Techspot review where the data point originated from? Anyone who can understand English text and can read a graph can see it and understand it. Like it or not, Tegra K1's GPU is in another league in the GFXBench benchmark (ie. more than 60% faster at the very least in a very apples to oranges comparison), so this is a silly argument to begin with.
no, I am pretty sure itsmy... was stating that you can't deduce any of this from an unreleased tk1 device and an AMD mullins prototype tablet...
There is nothing to deduce. Techspot clearly stated that they did all their published measurements on AC power, and when they switched to battery power, performance went down by ~ 10% due to less aggressive boosting of CPU/GPU operating frequencies. So on AC power in a relatively bulky 11.6" tablet, they measured 37fps in GFXBench T-Rex HD Offscreen with Mullins. On the other hand, on battery power in a relatively small 7.9" tablet, Xiaomi measured 60 fps in GFXBench T-Rex HD Offscreen with Tegra K1. That is a difference of more than 60% in favor of Tegra K1. Give credit where credit is due. Obviously TK1 is way ahead here, end of discussion.