Performance, Battery Life and Conclusion
The Prime is the first tablet on the North American market to offer the Tegra 3 system-on-a-chip. Popularly advertised as a quad-core chip, it’s actually a five-core chip consisting of four primary cores and a fifth low-power core. The purpose of the low-power core is to provide power for tasks don’t need Tegra 3’s full might, such as playing back audio with the display off. It’s a good choice that could potentially impact battery life, though it may be difficult to evaluate because battery life tests usually involve video or web surfing.
Clock speeds are up as well. While Tegra 2 in the original Transformer was clocked at 1 GHz, this new version is clocked at 1.3 GHz. There’s a minor dynamic overclocking feature as well, which allows one core to overclock itself to 1.4 GHz when the rest of the cores are at idle.
Tegra 3 also includes a revised GPU which increases the core count form 8 to 12 and, predictably resulting in a large theoretical performance increase.
At least on paper, Tegra 3 looks to be a huge leap forward. This could be deceiving, however. Improving performance by using multiple cores may be difficult, as anyone involved with PC hardware over the last seven years could tell you. In order for more cores to translate to better performance, the additional cores must be used. What do the benchmarks say? Let’s start with the browser benchmarks.
In our browser tests it appears as if the issue of poor multi-core efficiency is a problem. The results here don’t even seem to scale with the increase in clock speed, resulting in some rather measly gains compared to both the original Transformer and the Galaxy Tab 10.1. Only in Peacekeeper do we see an exception to the rule – there, the Prime’s score is much better than the Galaxy Tab or iPad 2.
If these scores are representative of the performance increase Tegra 3 offers in normal web use, then there’s not much improvement to be found. So, are they? In my opinion, yes. I didn’t notice anything about the Prime during web browsing (be it via the browser, or apps like YouTube) which was different from other modern Tegra 2 tablets. The iPad 2 still feels a bit quicker, particularly after the iOS 5 update.
Perhaps there is more to be found in the general and GPU-focused benchmarks.
In Quadrant we again see that there’s not a lot of performance improvement between the original Transformer and the new, Tegra 3 equipped model. Nenamark is a different story, however – there’s a substantial improvement there. While it appears Android is having trouble with making use of the multiple CPU cores, it certainly doesn’t have issue taking advantage of more GPU power.
As mentioned earlier, the Prime came with a number of games and tech demos designed to show off the new hardware. Though they mostly failed to be enjoyable, their graphics quality is undeniable. Nvidia has stated in press materials that they now offer console quality graphics, and it’s not much of an exaggeration. Yes, tablets are benefited by the fact that they have to render at low resolutions – but most current generation consoles titles are rendered at 720p and then upscaled for 1080p televisions.
Unfortunately, it’s not clear how this will translate into real-world results. Android games still tend to be created with a single quality standard. That’s beginning to change as devices become more diverse, but for now the number of games that can make use of Tegra 3’s power is small – and the number of games you’d actually want to play is even smaller.
Battery life is an area where the Prime could potentially show significant improvement. The combination of a slightly larger battery with a newer processor shows promise. To test it, I fired up YouTube and began my streaming video test. In this case I used a setting of 70% brightness without super IPS mode on, which should provide a good viewing experience in most lighting conditions. So, what happened?
Unfortuantely, it appears that battery life has taken a dive compared to the original Transformer. I was so surprised by this result that I re-tested battery life multiple time and at two different brightness settings to see high screen brightness impacted total run time. I was unable to drastically improve battery life.
I also decided to try out Peacekeeper’s new battery life benchmark. Although we don’t have data to compare it to other devices (since the new benchmark wasn’t out when we tested them) it can give us another point of data to consider both now and in the future.
This benchmark turned out to be less demanding than I expected, as the Prime was able to return results better than the streaming video test. Only barely, though – we gained 20 minutes at 70% brightness and 44 minutes at 30% brightness.
I should note that, while battery life did not impess me during these benchmarks, the Nvidia’s low-power fifth core does seem to kick in under certain scenarios. While listening to a podcast with the display off, for example, I found that battery life had dropped only a few percent over the course of an hour. In addtion, other review sites appear to have achieved much better results than these, but when I reviewed them to see why that might be, I found that most benchmarks were testing video without WiFi on.
Our battery life tests are more demanding, which drastically impacts life. If you use the full capabilities of the Prime, you will drain the battery with speed.
My feelings about the Prime are somewhat mixed. The design of the product is excellent. It’s thin, it’s light, and it’s beautiful. It feels like a product that is capable of fighting the iPad 2 without relying on a price handicap. ASUS has really outdone themselves here, and yet again shown that the company can – when they want to – deliver excellent aesthetic design.
Nvidia’s Tegra 3, on the other hand, fails to provide the sort of performance improvement that you’d expect considering a higher clock speed and a doubling of core count. Yes, there is an improvement – but at least on the CPU side, it is small. Graphics performance has been improved significantly, but still isn’t going to blow away the iPad 2. It may only be competitive rather than substantially quicker.
Is the problem a software issue? Probably. The tools for monitoring CPU use while running other apps are not good on Android, but from what I see of CPU performance, it seems likely that a lack of multi-core optimization is the issue. Ice Cream Sandwich could improve performance by offering better use of multiple cores, but until it’s actually available on the Prime, that is an open question.
The main problem with the Prime in my testing proved to be battery life. It seems to have actually lost out in comparison to the original Tegra 2 powered tablet, which is an unfortuante turn of events. The Prime will last a very long time if do not tax its full compliment of processors or its WiFi – but if you do engage WiFi and start to tax Tegra 3’s many cores, the battery doesn’t last as long as the original.
As such, the Prime finds itself in a similar place as the original Transformer. Yes, this is a very good tablet – and the best Android tablet for the money. However, the iPad 2 still appears to be the better choice. The OS is better, battery life is better, app selection is better, and performance is generally on par. The Prime is the closest any Android tablet has come to leaping the high bar set by the iPad 2, but it still can’t top it – and it won’t really have a chance of doing so unless battery life is improved and Android 4.0 is a major improvement over Honeycomb.