Archive for December, 2010

This blog originally posted on the Low Power Engineering Community 12/2/10.  http://chipdesignmag.com/lpd/absolute-power

This month, a little diversion from our recent deep dive into smart phones – I recently received in the mail my new MacBook Air!  First of all, kudos to the order fulfillment process.  I ordered my new baby on Friday.  It shipped from Shanghai on Saturday, and arrived at my door on Monday.  No special or delivery charges needed.  I had to order online in order to max out the memory, both RAM  (4GB) and SSD (128GB), as well as the processor (1.6GHz).  And it was worth the wait – this little (11.6’) machine runs really well considering its processor is somewhat dated, and it LOOKS GREAT!  At 2.2 pounds (most of which is batteries) and sporting a 16:9 screen, it’s a natural for video, surfing the web, or any other activities you’d normally do on your laptop.  Microsoft Office (2008) runs nicely, as does Adobe CS5, and VMWARE Fusion 3.  I even spent quite a bit of time running StarCraft II over Thanksgiving (as a gaming benchmark only, of course), and it did a reasonable job as long as the graphics settings were medium to low.

So it’s a wrap, right?  Not so fast…!  We all know from our recent investigations that we’ve gotten beyond just the specs on computers these days – GHz and GB are no longer the most meaningful metrics.  And while running real world applications can tell us something about the utility of computing devices, for portable electronics, it’s all about battery life, or power efficiency.  So I ran the Air through my recent power testing setup – determine the power efficiency of the machine for playing back our reference movie, “Star Trek”!

The current champion for personal movie watching has to be the iPad – it’s a great combination of function and form, with a beautiful display and fantastic battery life.  Watching “Star Trek” on these two machines was a great experience – both are very compelling as portable video players.  The Air has the definite advantage in screen size – it’s 16:9 aspect ratio was a better fit for the movie, resulting in a significantly larger picture even though the displays physically aren’t that much different in size (11.6” vs. 9.7”).  Both displays were nice and bright, but the iPad definitely wins for picture crispness and clarity.  The Air’s display is nice, but just did not seem to have the extra depth and contrast of the iPad display.

In terms of power efficiency, the Air carries a lot of baggage (for just watching a movie) – MacOS and the laptop platform is much more complicated than iOS, so we’d expect to see that overhead in the numbers.  And we sure do!  The iPad dissipated 6.2 Wh of energy to view Star Trek on maximum brightness and sound.  On the MacBook Air, the same setup used over 21 Wh of energy, nearly 3.5x!  Removing the display from the equation doesn’t help – with brightness and sound at minimum, the iPad uses 3.2 Wh of energy, compared to the Air’s 14.0 Wh.  That’s a 4.5x difference in energy efficiency.

So where does the new MacBook Air fit in today’s lineup of portable computing devices, and does it threaten the exploding market for the iPad and other tablets?  In my mind, it carves out a nice niche in the “ultraportable laptop” category, sitting between tablets, netbooks, and more traditional laptops.  With good performance, reasonable battery life, and style to spare, this category may well indicate the future of our laptops, absorbing the netbook category, as well as low-end traditional laptops.  But watch out for the growth of tablets in coming years.  Today’s tablets are attacking mobile computing from the standpoint of smart phones as opposed to laptops, and that philosophy seems to bring ultra-high efficiency, novel features, and slick packaging.  The Microsoft Office app can’t be far behind.

This blog originally posted on the Low Power Engineering Community 11/4/10.  http://chipdesignmag.com/lpd/absolute-power

As we look at power usage and efficiency of mobile devices, it seems only fair to expand our view beyond my personal stable of iOS devices.  This month, I’ve picked up a new Motorola DROID X phone, one of a spate of recently introduced smartphones looking to take a bite out of Apple’s iPhone business. 

And it’s an impressive piece of hardware, with similar specs to an iPhone 4.  A bright 4.3” display with 848×480 resolution is bigger, but lower resolution compared to the 3.5” 960×640 “retina display” on the iPhone 4.  The processors both run at 1 GHz, so both phones are quite capable, responding quickly to the multitouch interface, switching between applications and tasks with no problems.  I like the removable (and therefore upgradable) micro SD card on the DROID X, but mine was initially unrecognized in the device – I had to take out the battery and reseat the card for it to show up, and even now it’s a little flaky.  The quick hardware comparison round goes to Apple iPhone 4 – mostly for fit and finish – it’s just a plain sexy piece of technology.  The DROID X kind of reminded me of a Palm 3 (which I loved) when I first picked it up – a little “plasticky” compared to the evolved sensibility of the iPhone 4.

But perhaps the most important piece of hardware isn’t on the phone itself; it’s the cellular infrastructure that transforms these phones from novelty devices to modern day computing and communications devices.  My initial data on the “network” for each of these phones is a solid win on the DROID X side.  Verizon coverage is clearly superior to AT&T in the Palo Alto vicinity, including in my house (Palo Alto) and in my office (Mountain View).  It appears that with a good signal, the AT&T network may indeed be faster, but I’ll have some additional tests on that later.  Either way, from a user standpoint, consistent coverage definitely trumps network speed.  The network round easily goes to DROID X (Verizon).

On the software side, my new DROID X is running Android 2.2, the OS that has overtaken iOS on smartphones in only one year!  I’m new to Android, so not qualified yet to comment about daily usage, but my first impressions were that it seemed quite capable, on par with iOS.  However, I found a few subtle nuances in the user interface that were a little disconcerting – for example, scrolling the display had a slight delay, making it feel more like a simulation than the instant response on my iPhone, which really feels like you’re moving a piece of paper around.  And for the 2.2 rev of the OS, I found Android to be fairly unstable.  In addition to several apps crashing, my droid hung up quite a few times in the first couple of weeks, requiring reboot, and even got so far off into the weeds once that I had to take out the battery to reset it (can’t even DO that on my iPhone!).  More for stability than for features, iOS wins this round.

Applications software is difficult to rate; both platforms have many more apps available than I’ll ever try, but all of the requisite pieces seem to be there.  However, with regard to the overall user setup and interface, I have to say that I like the iTunes “all in one” idea much more than the “maybe-more-capable” but “definitely-more-hassle” setup on Android.  After seeing how much trouble it is to try to get everything set up on a new device, I’m much more forgiving (and appreciative) of some of iTunes’ restrictions.  Chalk another one up for Apple.

Killer apps on either device?  The one that sticks out on the DROID X is the nice integration in Android of voice activated mapping and directions.  Say “navigate to starbucks in palo alto” and it will do it!  I currently use a combination of google voice, google maps, and Navigon to do the same thing on my iPhone 4.  Nice job on this, Google.  On the Apple side, the killer app is simply that you get a piece of technology that lets you forget about technology;  ignoring the spotty coverage, it just works (someone should coin that phrase…)!  Next we’ll look at head-to-head power tests on the two platforms.

LATE-BREAKING FOOTNOTE: After my initial experience with the superior network coverage of the DROID X (Verizon), I went to the AT&T store and filed a complaint – and they gave me an “AT&T 3G MicroCell” device, which basically creates a mini cell site in your house!  I’ve been playing around with it, and I can now play the recent Star Trek movie on my iPhone 4 with just 2.8 Watt-hours of energy, streamed over 3G using the MicroCell (with a never-before-seen 5 bars of signal strength).  My previous tests consumed 3.4 Watt-hours in the same spot in my house (about 3 bars), and 5.3 Watt-hours in my poor-coverage (1 bar-ish) office.  And all of these measurements with brightness and volume at maximum.  As we postulated last time, signal reception may very well be the “smoking gun” of mysterious battery drain on these devices.  That’ll make running comparable power tests between the iPhone4 (AT&T) and DROID X (Verizon) a little trickier…

This blog originally posted on the Low Power Engineering Community 10/7/10.  http://chipdesignmag.com/lpd/absolute-power

Our cliffhanger from last time left us with the open question, “where does all of the energy in a 5.25Wh (watt-hour) iPhone 4 battery go during normal daily operation?”  Watching Star Trek as a test vehicle seemed amazingly efficient on iOS devices.  Our “Star Trek Inefficiency” metric (STI – energy used to watch Star Trek) showed STI=1.1Wh on an iPhone 4 in its most efficient configuration (airplane mode, display minimum, sound off), and STI=1.7Wh with the display and sound at maximum.  And further experiments demonstrate that energy used for sound via speaker, headphones, or Bluetooth are all small contributors to the energy equation.  So we’ve fingered the display.

But amazingly, that’s not the biggie!  As I’m sure you’re all screaming in your head by now (aren’t you?), we have conveniently ignored the contribution of the thing that makes today’s portable devices so compelling – COMMUNICATION!  Without the communications link, email is just a file of printouts, social networking is only possible within the range of your voice, and phone calls – well, when’s the last time you saw a phone booth?!  We are now completely dependent on wireless voice and data communications.  Remember when your parents used to complain to the phone company when there was some static on the phone lines?  Today, we pay nearly $100/month per person, and we live with service that cuts off 20% of all phone calls!

OK, rant finished – we’ve established that we can’t live WITHOUT it.  Now let’s see how much we’re paying to live WITH it – in today’s most important currency – ENERGY.

iPhone 4 STI=1.1Wh in its most efficient mode.  STI=1.7Wh with brightness and sound at maximum (it’s a great movie player).  Now, instead of playing the Star Trek movie on the device, let’s stream the movie via 3G!  The iPhone 4 STI with Star Trek streamed over 3G, STI=3.4Wh.  Wow!  Streamed (at a much lower resolution) via Netflix, Star Trek now requires twice as much energy as the full brightness case!  I ran this experiment at my house in Palo Alto, where my reception is pretty consistently between 3-4 bars.  Remember, battery capacity is 5.25Wh, so we’ve gone from 1/3 of the battery capacity to watch Star Trek locally in high resolution to 2/3 of the battery capacity to watch it streamed in low resolution.

And for my final trick, I wanted to find the perfect spot to rerun this test – 3G reception good enough to maintain the stream of data, but otherwise right on the edge.  A place of great 3G signal suffering.  Not a dead zone, but something like eternal reception purgatory.  No bars…one bar…two bars…one bar…  Eureka!  On my desk in my office!!!  The perfect spot (and I’m SURE one of only a few spots on earth) where you can be tempted, enraged, encouraged, and frustrated all while trying to read your email.  In this “mystery spot”, where good 3G reception and evil 3G reception wrestle to the end of time, the STI on my iPhone 4 came out to be 5.3Wh!!  Woah!  You’re probably thinking, “Hey, I thought he said the battery capacity on the iPhone 4 was only 5.25Wh?”  And you’d be correct – in this magical spot (where I sit everyday), I can drain a fully charged iPhone 4 battery in less than two hours, and not even see Spock do his little telepathic speech at the end!

The revised list of ways to improve your iPhone 4 battery life:

3. Forget everything else on the lists you read on the internet

2. Turn down the brightness on your display

1. Improve your reception

It turns out that “improving your reception” isn’t so easy to do – you could be promoted and move to the corner office, sell your house and choose your new one by “closely watching the little bars”, or for no money and little fuss, simply get a case for your phone!  Remember iPhone 4 “antenna-gate”, resulting in the Apple free case program?  It turns out that using a case for your phone not only avoids the “death grip” problem, but also moves your hand (which is mostly water) a little bit away from your phone, slightly improving antenna performance and reception in both cases.  The result is a one or two-bar improvement in signal reception, especially in areas of generally weak reception.  And we’ve already seen that one or two bars of reception can have a profound impact on battery life – in our case, a jump in the STI from 3.4Wh to 5.3Wh just based on those one or two little bars.  So, for the purposes of streaming Star Trek over 3G, that free case may just contain a supplemental 1.9Wh battery!  But don’t let Apple know, or the prices will be going up…

And yes, I got my case through the free case program, but I’m not using it.  Why not?  I just don’t like the way it looks. (Go figure!!!)