MacBook Air HDD and SSD battery benchmarks
The Wall Street Journal's Walt Mossberg (SSD) and AnandTech's Anand Lal Shimpi (HDD) have each issued reports on the battery performance of Apple's new MacBook Air models after putting the notebooks through some extensive real-world tests.
For his part, Mossberg pit both 1.6GHz MacBook Air SSD and a 1.6GHz MacBook Air HHD against each other by turning off all power-saving software, setting screen brightness to maximum, turning on Wi-Fi, and playing an endless loop of music.
"In this test, the SSD made little difference in the MacBook Air and, in fairness, Apple is making no claims of any significant battery-life gains on its SSD model," he wrote. "The SSD MacBook gave me just five more minutes of battery life. Apple says this is because its hard-drive model already uses a very low-power drive."
Next, Mossberg ran a system boot test and found that the SSD version booted up from a cold start, and rebooted with several programs running, about 40 percent faster than the HDD model.
"But the gain isn’t as impressive as it seems because even the hard-drive versions of the MacBook Air booted up in under a minute and rebooted in just a little over a minute," he explained.
Overall, Mossberg dubbed the advent of notebook-grade SSDs a "promising improvement over the hard drive," but said "now is not the time for most users to buy it."
Over at AnandTech, Lal Shimpi focused his battery benchmarking efforts off a 1.8GHz HDD-based MacBook Air, as he set out to discount Apple's 'laughable' claims of 5 hours battery life.
A wireless web browsing test used an 802.11n connection to browse a series of 20 web pages of varying size, spending 20 seconds on each page. The test continued in a loop while playing MP3s in iTunes, yielding battery life of 4.27 hours.
A second test simply played a copy of the DVD Blood Diamond -- which had been ripped to the hard drive -- in a loop until the battery ran out, yielding battery life of 3.42 hours.
In a final test emulating a multitasking workload, 10GB of files were downloaded in the background while simultaneously running the aforementioned web browsing test, and playing back in QuickTime the first two episodes of Firefly encoded in a 480p XviD format. Battery life during this test lasted just 2.43 hours.
Source: AnandTech
For each of the three tests, the MacBook Air's display brightness was set at 9 blocks (just over 50 percent), and the system was set to never shut off its display and never go to sleep, although the hard drive was allowed to spin down when possible.
Apple's 5 hour claim is laughable but not as much as I expected. If I wanted to I suspect I could hit 5 hours by making the web browsing test less stressful, but my focus was on real world usage scenarios, not proving Apple correct," Lal Shimpi concluded. "Regardless, 4 hours and 16 minutes doing what I consider to be the intended usage model of the Air is respectable. It's not great, but it's not terrible either."
AppleInsider is presently putting its own MacBook Air SSD review model through the ringer, and will have some reflections soon. Meanwhile, readers are welcome to pose questions for Prince -- who will be conducting the reviews -- and make their own benchmark suggestions or requests in the comments section of this article.
Earlier this week, AppleInsider posted its own review of the 1.6GHz HDD-based MacBook Air.
For his part, Mossberg pit both 1.6GHz MacBook Air SSD and a 1.6GHz MacBook Air HHD against each other by turning off all power-saving software, setting screen brightness to maximum, turning on Wi-Fi, and playing an endless loop of music.
"In this test, the SSD made little difference in the MacBook Air and, in fairness, Apple is making no claims of any significant battery-life gains on its SSD model," he wrote. "The SSD MacBook gave me just five more minutes of battery life. Apple says this is because its hard-drive model already uses a very low-power drive."
Next, Mossberg ran a system boot test and found that the SSD version booted up from a cold start, and rebooted with several programs running, about 40 percent faster than the HDD model.
"But the gain isn’t as impressive as it seems because even the hard-drive versions of the MacBook Air booted up in under a minute and rebooted in just a little over a minute," he explained.
Overall, Mossberg dubbed the advent of notebook-grade SSDs a "promising improvement over the hard drive," but said "now is not the time for most users to buy it."
Over at AnandTech, Lal Shimpi focused his battery benchmarking efforts off a 1.8GHz HDD-based MacBook Air, as he set out to discount Apple's 'laughable' claims of 5 hours battery life.
A wireless web browsing test used an 802.11n connection to browse a series of 20 web pages of varying size, spending 20 seconds on each page. The test continued in a loop while playing MP3s in iTunes, yielding battery life of 4.27 hours.
A second test simply played a copy of the DVD Blood Diamond -- which had been ripped to the hard drive -- in a loop until the battery ran out, yielding battery life of 3.42 hours.
In a final test emulating a multitasking workload, 10GB of files were downloaded in the background while simultaneously running the aforementioned web browsing test, and playing back in QuickTime the first two episodes of Firefly encoded in a 480p XviD format. Battery life during this test lasted just 2.43 hours.
Source: AnandTech
For each of the three tests, the MacBook Air's display brightness was set at 9 blocks (just over 50 percent), and the system was set to never shut off its display and never go to sleep, although the hard drive was allowed to spin down when possible.
Apple's 5 hour claim is laughable but not as much as I expected. If I wanted to I suspect I could hit 5 hours by making the web browsing test less stressful, but my focus was on real world usage scenarios, not proving Apple correct," Lal Shimpi concluded. "Regardless, 4 hours and 16 minutes doing what I consider to be the intended usage model of the Air is respectable. It's not great, but it's not terrible either."
AppleInsider is presently putting its own MacBook Air SSD review model through the ringer, and will have some reflections soon. Meanwhile, readers are welcome to pose questions for Prince -- who will be conducting the reviews -- and make their own benchmark suggestions or requests in the comments section of this article.
Earlier this week, AppleInsider posted its own review of the 1.6GHz HDD-based MacBook Air.
Comments
W00t First Post...! So actually, these figures are pretty impressive. I thought 5 hours means 3 hours. But this is nice info for my customers to show real-world battery benchmarks. Bring it on.
For the last 4 years Apple has been pretyt accurate with it's battery numbers. I wonder if he turned off iTunes and only searched websites if he could get 5 hours out of it. That is what I do mostly when abroad and with internet.
I'm glad AnandTech stepped up after after the ArsTechnia post. THose ArsTechnia findings of 2.5 hours with the display down to the lowest setting seemed to awful to be true. It was more fodder for the Anti-Apple collation.
Well, as irrational as it is, the thought of having a computer with no moving parts will still make me go for the SSD if/when I have the cash. Call me an Apple whore!
It's not completely without moving parts. There's a fan in there.
Who actually plays movies while downloading large files and listening to iTunes (whether plugged in or not)? Probably no-one. My typical use of my MacBook Pro is on an airplane, using Powerpoint or Word, maybe listening to iTunes (but more likely my iPod), with the screen brightness at only one square (which works just fine on a plane). I'd be much more interested to know how long the battery lasts under those conditions.
It's not that unrealistic. They're finding what may be the upper and lower bounds of the battery life for a good variety of use.
The Wall Street Journal's Walt Mossberg (SSD) and AnandTech's Anand Lal Shimpi (HDD) have each issued reports on the battery performance of Apple's new MacBook Air models after putting the notebooks through some extensive real-world tests.][/url][/c]
Would love to see how it stacks up against Sony. Their long-life, laptop battery-life test results are determined by playing MP3s only.
I'd say Apple is pretty well vindicated by this test. They got 4.5 hours with continuous WiFi use and continuous MP3 playing (iTunes turns off the disk cache, so this amounts to continuous hard drive access). It's not hard to see how that would correspond to 5 hours of typical usage.
I don't understand how 4.27 hours in the article jumped to 4.5 hours in your post.
Your definition of continuous hard drive access could be pretty sketchy. iTunes could just read the entire song (or more) in and not hit the hard drive more than once every five minutes. I just don't see how it necessarily scales to five hours under "normal" use. Both types of hard drives consume very little power anyway.
Would love to see how it stacks up against Sony. Their long-life, laptop battery-life test results are determined by playing MP3s only.
IT wouldn't be a good real world test for this machine but that is an excellent idea of a straight on comparative test.
Would love to see how it stacks up against Sony. Their long-life, laptop battery-life test results are determined by playing MP3s only.
Then it's good that there are third parties that test this sort of thing:
Six hours 43 minutes:
http://www.notebookreview.com/default.asp?newsID=3956
Four to six hours:
http://www.geek.com/review-sony-vaio-tz-notebook/
It's not that unrealistic. They're finding what may be the upper and lower bounds of the battery life for a good variety of use.
For me the argument is the lower bound. I would not be running my screen at a six point brightness, I would do a three or four, and I can't browse the internet while listening to music, it's too distracting and not relevant to a low-bound testing anyway. Moreover, when browsing the internet, I could stay on as much as 10 minutes or more on some pages, not 20 secs.
The lower bound benchmark should've looked more like this:
- Screen brightness 4.
- Script to browse internet at 10 min per page.
- Only Mail and Safari running, checking email every 5 mins.
yeah.. that looks more like a low bound test.
IT wouldn't be a good real world test for this machine but that is an excellent idea of a straight on comparative test.
Exactly. The Sonys, Toshibas, etc., claim in print that their ultralight batteries' life span is up to 10 hours. However, unless one looks at the references closely in their respective manuals under the 'legal' disclosure section does one find that the tests were done by measuring how long they ran playing MP3 files only. (I would gather that the lids were down)
I would just like to see how long the Macbook Air would run under the same protocol. Then perhaps I could make up my mind whether I buy a Macbook Air or a comparable number of iPod Touches.
For me the argument is the lower bound. I would not be running my screen at a six point brightness, I would do a three or four, and I can't browse the internet while listening to music, it's too distracting and not relevant to a low-bound testing anyway. Moreover, when browsing the internet, I could stay on as much as 10 minutes or more on some pages, not 20 secs.
The lower bound benchmark should've looked more like this:
- Screen brightness 4.
- Script to browse internet at 10 min per page.
- Only Mail and Safari running, checking email every 5 mins.
yeah.. that looks more like a low bound test.
Maybe it's a lower bound for you, but it's not necessarily a lower bound for the machine. I'm talking about the machine, not for a given user.
A ten minute average per page is a long time, I'm very skeptical of that. I don't doubt there are pages worth that, but I doubt that's a realistic average, many pages might be rejected before settling on one to read, or you follow a few links to find the one article that you want. One figure I saw from Neilsen testing was 43 seconds spent on a page average. So mybe the Anand test is a tad aggressive, yours is excessively lax.
That being said, LED backlit displays also do not offer a significant power savings over tubes used in LCDs. Again, look at the paper specs for LED vs. tube displays. There is more power savings of a LED backlit display than a SSD. But overall, it is not a giant leap in power savings. The true benefit of LED in my opinion is brighter light and longer life (and less hazardous materials for the environment).
There is a tremendous opportunity for battery performance improvement using nanotechnology, i.e. increasing the surface area of the capacitive components of the battery. We will see super fast charging (an order of magnitude improvement) and longer life due to greater efficiency. Hopefully we will start seeing some of these batteries hit the market soon.
- Screen at a single bar brightness
- Wifi on, connected to a 802.11n network
- Leopard set not to sleep (though the hard drive / SSD can)
- Nothing else: just leave it on, idling away to see the maximum reach
2. Offline Writing Test- Same as test 1 but with no wifi either
- Use a script to enter text into TextEdit, Word or Pages* at a ~20 wpm rate!
- That is all
3. Online Browsing Test- Same as test one but using Safari to load a page every 5 minutes, a quarter maybe containing flash (irritating ads / embedded video)
- That is all
Note that these are all pretty low power use. That's what mobile users like me have been doing with our Macs all these years, getting 4 hours or so out of a battery. My 12" PowerBook pretty much gets 2 and 3 (but with me hitting the keys) most days I use it.*Although on that old PowerPC I do notice Pages is a real power hog!
Mobile audio is definitely a job for the iPod in your pocket, instead of the Mac on your knee. Portable video meanwhile is a better question … iPod touch / iPhone comes to mind, but whatever don't spin a disc too!
As for the low screen brightness: that's one thing myself and Ars' Jacqui Cheng are agreed on. I do that even when hooked to mains just for comfort's sake.
FWIW, I sat down in bed last night at exactly 10pm and started working.
Lots of email. Lots of web browsing.
No music, No movies.
All via Wifi of course.
No bluetooth.
Monitor set to a comfortable brightness but not overly bright.
Of course the keyboard was backlit.
At exactly 1:30 am I was down to the warning and 6 minutes of battery life. This is my first mac so I don't know how long I really could have worked once it read 6 minutes but I do know that 2 minutes later it read 4 minutes.
I love this machine and am pretty happy with Leopard and Mac apps in general but I was fairly disappointed that very basic use drained the alleged 5 hour battery in 3.5 hours. That's 70% of the promised battery, I expect if I was watching a movie on a plane or listening to music I'd be hard-pressed to get 2,5 hours.
Even worse, if 6 months or so the battery will probably offer a lot less as it gets tired.
Anyone else have any real-world anecdotal evidence to battery life?