Samsung Galaxy Nexus & Ice Cream Sandwich Review
by Brian Klug & Anand Lal Shimpi on January 18, 2012 1:34 PM ESTThe SoC: TI's OMAP 4460
The launch platform for Ice Cream Sandwich was TI's OMAP 4460. Unlike previous Android releases however, it seems that other SoCs will see their ICS ports done in a much quicker manner. It took a very long time for Honeycomb to be ported to other SoCs, whereas a number of companies have already demonstrated ICS running on their hardware (e.g. Intel, NVIDIA). If this is the case going forward, the launch vehicle for a new Android version may not mean what it used to.
The OMAP 4460 is a fairly standard, yet full featured dual-core ARM Cortex A9 SoC. You get two A9 cores complete with MPE (NEON support), behind a shared 1MB L2 cache. The SoC features two 32-bit LPDDR2 memory channels as well. The GPU is provided by Imagination Technologies in the form of a PowerVR SGX 540.
Max clocks for the OMAP 4460 are 1.5GHz for the CPUs and 384MHz for the GPU. As with all SoCs, all final clocks are OEM customizable to hit their desired point on the performance/battery life curve. Google and Samsung settled on 1.2GHz for the cores and 307MHz for the GPU, both exactly 80% of the OMAP 4460's max frequencies. Sprint recently announced its Galaxy Nexus would run at 1.5GHz. It's quite possible that we'll see a jump in GPU clocks there as well since the two may run in lockstep.
From a CPU standpoint the 4460 is competitive with pretty much everything else on the market (A5, Exynos, Tegra 2, Snapdragon S3). The 4460 does have more memory bandwidth than Tegra 2, Tegra 3 and Snapdragon, but it's comparable to Apple's A5 and Samsung's Exynos 4210. It's the GPU that's a bit dated at this point; the PowerVR SGX 540 typically delivers Tegra 2-class performance. A quick look at GLBenchmark and Basemark results echoes our findings:
At 720p, which happens to be the GN's native resolution, the OMAP 4460 is much faster than Tegra 2. It's also important to note just how much faster Tegra 3's GPU is by comparison.
I understand why Google didn't wait for a Krait based SoC, however I don't believe the OMAP 4460 was the best bet given the launch timeframe of the Galaxy Nexus. Based on performance alone, Google should have picked Tegra 3 as the launch platform for ICS. GPU performance is much better than the SGX 540 and there's comparable CPU performance. It's possible that Google needed the memory bandwidth offered by OMAP 4, but we'll find out for sure soon enough as the first Tegra 3 device (ASUS' TF Prime) is slated to get ICS this week.
I'm also less concerned about power consumption being an issue since NVIDIA added full power gating to all of the cores in Tegra 3. With a conservative enough power profile Google could have guaranteed battery life similar to OMAP 4460 out of Tegra 3.
I get the feeling that Google wasn't very pleased with NVIDIA after Honeycomb and chose to work with TI this time around for reasons other than absolute performance. If it weren't for the fact that Tegra 3 and other SoCs appear to be getting ICS in fairly short form I'd be more upset over this decision. To be honest, the choice of SoC simply hurts the Galaxy Nexus as a phone. If I were you, I'd wait for a Krait based device.
Facebook
Twitter
RSS
185 Comments
View All Comments
Tujan - Wednesday, January 18, 2012 - link
Concerning the youtube video shown about the android smartphone. No misconception here,but is it the intention of the video to put your thumb so much at the forefront of the small phone,somehow toggling between 'both eyes' of the viewer. ? Is that the intention of the video ?Pure punn intended that is an awesome big speaker you are speaking to. And true even though the video was most likely rendered in 'full part',with both video and audio attached,I cannot help the feeling that there IS SOME LATENCY between my chinese made glasses and the finish of my lcd.
The fact that you are discussing a radio device,and I am utilizing a later viewing of it over a wired internet connection. Does not diminish the fact that your radio devices capability of facial recognition,and my lcds display of it is no substitute for taking x-rays if you actually need to.
I see that you are descript in functioning your arms across the whole of the screen at the making of your video. And there is that very large meter between the preposition of the distance of the audio device,and the radio device that is to conclude that preposition.
Punn accepted there is certainly some latency there that is perfectly conceptual. Between the foreground,and the background. And the autonomic acceptance of my viewing it.
You notice that at times as a forefront,you have a wide screen rendering. Then at other times there is the focus 'in'. The difference in doing so is the focal point of my comment in that subject of its latency.
And that truly the speaking IS a separate distinction of a Microphone. Than that of a speaker,and the screen displayed. Your being behind it shouldn't be misconstrued of what my comment is coordinating to account to. Since obviously the latency between my glasses and what I see on the screen at my viewing of it is of no consequence to your creation of it.
Mentioning that relationally you cannot change the environment around you no more than I can make your video for you. Perhaps someone will recognize this.
And thanks.
nsnsmj - Wednesday, January 18, 2012 - link
I always enjoy how detailed the reviews are.BitGambit - Wednesday, January 18, 2012 - link
I love this phone, but I have owned 2 different Nexii with the "Mura" screen issue. The first one I owned was glaringly obvious, but the second one less so, but it's still there. I was not able to exchange it for the second time, because the defect wasn't apparent enough to warrant an exchange, explains the Verizon employee. It pains me because having a good screen is important to me and I was looking forward to release of the Galaxy Nexus. I'm absolutely jealous of those who have a Galaxy Nexus with immaculate AMOLED screens.crankerchick - Wednesday, January 18, 2012 - link
I haven't had a chance to read the whole review, but I'm happy to see it. I've been checking everyday thinking I must have missed the review. It's great to see time and care being put into the review, as always. The video at the bottom was also very insightful and hits right in the points if why I prefer Android and the evolution it has seen, over iOS when it comes to my mobile devices. ICS flies in my Galaxy Nexus and my XOOM! I don't ever see 25 Mbps on my Nexus though, or all the bars for that matter. Maybe I have a signal issue? :-pflomt - Thursday, January 19, 2012 - link
I see you have the iphone 4s getting 9.85 hours of web browsing. Do you receive a phone from Apple to test, or do you go to the store and buy one?The reason I am asking is I have a 4s and I can tell you mine, and the people I know that have one are lucky to get 9.85 hours of battery life with the phone sitting on the nigh stand.
Thanks for the great reviews.
doobydoo - Thursday, January 19, 2012 - link
I think all the figures Anandtech post are their own measured statistics.I have a 4s and I, and the people I know that have one, are amazed by how long the battery life lasts, both in general and when web browsing.
When sitting on the night stand for 9.85 hours a very small percentage of battery life is depleted.
flomt - Thursday, January 19, 2012 - link
I know 4 other people with a 4s, none of them can last 24 hours with very light usage. 9 hours of web browsing is not even kinda of possible. All are on the most current release 5.0.1Reading the Apple forums, I am not the only one. Do the phones with poor battery life all originate in a different factory than the ones that last a long time? have a different version of the radio?
Apple is just denying that they have a problem now and I can tell you that they really do have a serious problem with some phones.
tipoo - Thursday, January 19, 2012 - link
Are you running push e-mail by any chance? That kills idle battery life.flomt - Thursday, January 19, 2012 - link
Nope, manual sync only.tom5 - Thursday, January 19, 2012 - link
You state in the article here that the camera sensor is a NOT-back-illuminated S5K4E1G sensor and Chipworks in the teardown article states that Galaxy Nexus is using "S5K4E5YA 5 Mp, 1.4 µm pixel pitch back illuminated CMOS image sensor":http://goo.gl/gvIWV
Who is right?