ArsTechnica

Gear & Gadgets / Product News & Reviews

Divine intervention: Google's Nexus 7 is a fantastic $200 tablet

Like Microsoft, Google shows its own partners how one builds a tablet.

Performance

The Nexus 7 is packing a 1.3GHz Tegra 3 quad-core processor and 1GB of RAM, a configuration it inherited from its spiritual predecessor, the MeMo 370. We don't need to point out that these specs are quite impressive given the price point. It was even a surprise to those who attended CES. The usual caveats apply (more cores are not necessarily better than fewer, etc.), but the Nexus 7 performs shockingly well in benchmarks, even compared to the iPad 2 and "3." It's almost as if Google thought ahead and tweaked the Nexus 7 to blow the benchmarks it thought we would use out of the water.

Throughout this section, we do compare the Nexus 7 to the last two generations of iPads (see the reviews for the iPad 2 and "the new iPad"). We also compare a couple of notable Android tablets from the recent past just to show how much the landscape has evolved in a matter of months. Finally, we compare it to the Kindle Fire in the only statistical measure it is capable of (the browser-based SunSpider).

We compare at this level not necessarily because we consider the Nexus 7 and iPad to be equals. The cheapest iPad is still twice the price of the base Nexus 7 at $399 after all (though bump the base Nexus 7 up to the same storage size as the base iPad, and only $150 separates them). While the Nexus 7 is definitely gunning for the Kindle Fire in terms of size, setup, and approach, we'd actually say it's a closer relative of the iPad, the homo floresiensis to the iPad's homo sapiens. We mean no disrespect by that; we don't foresee the Nexus 7's extinction in the current tablet climate.

Linpack performance scores (longer bars are better).

Up front, Linpack shows that more cores will not necessarily take you more places. In single-threaded processes, the Nexus 7 comes in well under the iPads, but handily defeats the Samsung Galaxy Tab 8.9 we reviewed last fall. In multi-threaded processes, the Nexus 7's four cores scale it right up to the league of the iPads. The Tab 8.9 is picking MFLOP daisies in the multi-threaded outfield.

SunSpider 0.9.1 (shorter bars are better). Galaxy Tab 8.9 scores culled from Anandtech, Galaxy Tab 7.7 scores from Galaxy Tab Life

In the JavaScript benchmark SunSpider 0.9.1, the Nexus 7 is bested only by the Samsung Galaxy Tab 7.7 released in fall 2011 (which, according to several accounts, seemed to perform freakishly well in only this metric while being thoroughly unimpressive in other measures like Linpack). The Nexus 7 bests the iPads 2 and "3" by only a trivial amount on the scale of SunSpider's measurements, while it handily defeats the Kindle Fire. The usual caveats apply to using a JavaScript engine to measure the efficacy of a browser, as we wrote in the iPad 2 review:

A poor JavaScript implementation on an insanely fast device can mask the true performance of the platform. Alternatively, a great JavaScript implementation can make up for a slightly slower system. More and more these days, the major JavaScript implementations are converging and borrowing ideas from each other. JavaScript tests are inherently single-threaded due to the nature of browsers. The SunSpider avoids many micro-benchmarks like we see in applications like GeekBench, and it attempts to benchmark processor-intensive tasks that might actually be found in a real Web application.

It's further worth noting that Android and iOS use two different JavaScript engines: iOS uses Apple's Nitro, while Android uses Google's V8. There are some differences in the way they handle Web content.

GLBenchmark scores (longer bars are better). Android devices are running v2.1.4, iOS devices are running v2.1.3 (the highest version available). Scores for the Galaxy Tabs 7.7 and 8.9 are taken from the GLBenchmark website.

As for GLBenchmark, the Nexus continued to impress with its 12-core NVIDIA GPU, going frame-for-frame-per-second with the iPad 3. (The graphs may seem strange in that the iPad 2 outscores the iPad 3. Given that the iPad 3's beefier hardware seems to exist entirely to support the hard work required to pump out four times as many pixels as the iPad 2, this is not an outrageous result).

Normally we run both the "high" and "standard" versions of GLBenchmark's Egypt and Pro tests, but when we ran the high version on the Nexus 7, the results read only "FSAA [full-scene anti-aliasing] is not supported" on the Nexus 7. We've never seen this before in GLBenchmark on an Android device or otherwise, but FSAA seems to be used to make a small improvement to graphics at a high computational cost. We're not entirely surprised it's disabled.

Geekbench scores (longer bars are better).

For an all-around benchmark, we used Geekbench 2, and were surprised to see the Nexus 7 punish this metric as well (for reference, a mid–2010 MacBook Pro scores around 5000 on Geekbench). While the iPads 2 and 3 roughly track with one another, the Nexus 7 speeds ahead. Its only notable weakness in Geekbench was its stream performance, where it received scores in the 260–290 range (the iPads generally clocked in around 330–340). The stream tests are a measure of memory bandwidth, and more memory bandwidth helps when working with large amounts of data (editing big photos or video, for instance). If you're drawn to tablets for their content creation abilities, the Nexus 7 is not as good a choice as an iPad.

It's important to remember that the Nexus 7 is, well, a 7-inch tablet—much less is asked of its internals than the iPad 3, and its hardware packs slightly more of a punch than the nigh-16-month-old iPad 2. Its small size is to its advantage, efficacy-wise, but size still comes with some downsides: conventional magazine pages and comic book pages are hard to read. You can't get much distance from it when watching a movie. It's the opposite thesis of the iPad 3: rather than ratcheting up the internals simply to serve a crazy-high-res screen, a smaller (still pretty high-res) screen lets very good internals shine.

Still, the Nexus 7's scores all around are good. Good to the point that you worry it's like that kid in high school who could recite his physics textbook backwards but was completely stumped at how to interpret experimental results. But so far the Nexus 7 performs well both experimentally and theoretically; it's really that star student on the football team (from the wrong side of the tracks!) who you suspect has a mild-to-moderate dependence on Adderall.

In practice, aside from the occasional keyboard lag we described earlier, the Nexus 7's performance is blisteringly fast. There's little hesitation from the OS at any turn. The device is responsive and quick, and it's a pleasure to interact with. Apps open quickly and perform very well. Swipes, taps, and pinch zooms are all accepted with no hesitation.

How the apps look is another matter. While most apps scale up pretty well to the 7-inch size, there are still some noticeable fuzzy edges here and there. The Android tablet market has provided little incentive for developers to scale up the graphics on their apps for larger screens; 7-inch tablets even less so, as regular smartphones edge closer to that size (witness the strange, and strangely popular, 5.3-inch Galaxy Note). The difference in graphics between the Nexus 7 and larger smartphones is small but visible.

iOS developers felt compelled to start offering new, scaled-up versions of their apps for the iPad because the iPhone and iPad screens are very disparate in size—an app made for a 3.5-inch screen drawn at 10.1 inches looks awful. But an app designed for a 4- or 4.5-inch screen that needs to scale up to 7 inches… we expect developers may be more inclined to let that difference ride. That will be to the detriment of the Nexus 7.

You May Also Like