The Best Battery Life I’ve Ever Seen

To find out how well the new lithium polymer battery does I ran my usual suite of Mac battery life tests. First up was my wireless web browsing test:

The wireless web browsing test uses the 802.11n connection to browse a series of 20 web pages varying in size, spending 20 seconds on each page (I timed how long it takes me to read a page on Digg and came up with 36 seconds; I standardized on 20 seconds for the test to make things a little more stressful). The test continues to loop all while playing MP3s in iTunes.

This is an extremely light test as none of the web pages have any flash ads, but it’s a valid test of very light wireless usage.

Eight, freakin, hours. I couldn't believe it. In my lightest test, the new 15-inch MacBook Pro lasted eight hours and eight minutes. That's with the screen at half brightness (completely usable) and no funny optimizations. The notebook is just playing music and surfing through a lot of my old reviews. There's no way this could be right. Maybe my test was too light?

I threw together another test just to make sure. The key flaw in my initial wireless web browsing test is that it none of web pages have any Flash on them. While constantly loading web pages will ensure the CPU can't go into deep sleep, Flash on the pages would make sure that the CPU utilization remains higher at all times. The next test I put together was this:

I strung together 8 reviews on AnandTech and put them each on a single page, images and all. I then scoured the web for big, animated Flash ads and added anywhere from 1 - 4 ads per page; all Flash. Each page is designed to forward to the next after 10 seconds and the loop continues indefinitely. On each machine I opened three Safari windows and pointed them at the first page in the sequence. In the background, once more, I had iTunes playing MP3s.

I found that CPU utilization varied from 5 - 35% during this test, which is about what I saw when I was actually surfing the web myself. The addition of Flash should make it more stressful, but it's still a fairly light usage test. My original web browsing test got us 8 hours, so what about this new one?

  MacBook Pro 2009 MacBook Pro Late 2008
Wireless Web Browsing w/ Flash 6.48 hours 3.28 hours

 

Six and a half hours, out of a 5.5 lbs notebook. For comparison, the older MacBook Pro could only manage 3 hours and 17 minutes in the same test. The new notebook lasted almost twice as long. Mathematically, this doesn't make sense. There's only a 46% increase in battery capacity, there shouldn’t be a ~100% increase in battery life...ever.

While the original web browsing test was using data from my original unibody MacBook Pro review, this second web test used a brand new MacBook Pro (purchased just weeks before this week's MacBook Pro announcement). The two notebooks had the same amount of memory (4GB), the older MacBook Pro had a slower CPU (2.4GHz vs. 2.53GHz) and a 7200RPM hard drive but the differences shouldn’t account for an extra 54% increase in battery life.

Apple must have done more than just increase battery capacity in the new MacBook Pro. My third test continues to support my findings. This is my heavy workload benchmark.

For this benchmark I'm downloading 10GB worth of files from the net (constant writes to the drive), browsing the web (same test as the first one) and watching the first two episodes of Firefly encoded in a 480p XviD format (Quicktime is set to loop the content until the system dies).

The older MacBook Pro managed 3.25 hours in this test. The new one? Just under 5:

That's a 51% improvement in battery life. It's close enough to the max theoretical 46% improvement for me to think that the significant gains in wireless web browsing are due to improvements in idle power optimizations. It's possible that all of the components in the new MacBook Pro have been optimized for lower voltages at idle.

The battery tests are repeatable however. I saw anywhere from a 50 - 100% improvement in battery life over the old MacBook Pro. Given the increase in battery capacity alone, you should see no less than a 46% increase in battery life. Exactly what is accounting for the expanded life above and beyond that, I'm not sure.

Either way, Apple's 7 hour claim is well within reason. For light workloads, even on WiFi, you can easily expect 6.5 - 8 hours out of the new 15-inch MBP. As I write this article on that very system I'm told that I have nearly 8.5 hours left on my charge. If you do a lot of writing on your notebook, the new MBP is exactly what you'll want; it will easily last you on a cross-country flight if you need to get work done.


I think I've just found my new writer's companion

My heaviest workload delivered just under 5 hours of battery life, a figure that the old MBP could only attain while running my lightest workload. This thing rocks.

I also have to commend Apple for delivering realistic battery life specs on its laptop. While 7 hours definitely involves a light workload, it is more than attainable as I've shown in the tests above.

A quick search shows that even Dell's Studio 15 only offers a battery rating of up to 5.5 hours. It looks like, once again, other notebook makers will have to play catch up to Apple in this department.

Other Hardware Changes Lower Power Consumption = Smaller Power Bricks
POST A COMMENT

113 Comments

View All Comments

  • ltcommanderdata - Friday, June 12, 2009 - link

    When Anand breaks out the results in the table for a head-to-head comparison he is comparing the new model against the Late 2008 Unibody which uses a 2.4GHz P8600 processor with a 25W TDP. So the TDP is unchanged.

    I believe the 2008 MacBook Pro (non-Late) in the charts with a 2.5GHz processor is a 800MHz Penryn, the T9300. I don't believe he uses this as his direct comparison since that MacBook Pro is 2 generations back.
    Reply
  • PlasmaBomb - Friday, June 12, 2009 - link

    Sorry the processor question was only a musing on my part, since on page 3 it states the cache of the current MBP but doesn't for the early '09. I merely wondered if the early 09 was a 35W TDP piece, now being changed to a 25W TDP piece.

    I don't have a mac and am purely interested in the tech :)

    The main question was really supposed to be "Could the absence of the Nvidia 9600 GT be partly responsible for the increased battery life?"

    Another question would be why didn't they update the dedicated graphics to the 55nm Nvidia GT 130M (basically a higher clocked 9600GT). Although I think Anand really answered that with - not many people use the graphics, so it might be an unjustifiable expense to apple :(
    Reply
  • PlasmaBomb - Friday, June 12, 2009 - link

    Sorry typing too quickly the processor number should be T9550 :( Reply

Log in

Don't have an account? Sign up now