Review and migration to the MacBook Air 2.0

The following is my lengthy story and review of the new MacBook Air 2.0. I get a bit long winded at times, but what geek doesn’t right? right?

Like most Apple users & developers I watched the recent announcements of new MacBook (MB), MacBook Pro (MBP), and updated MacBook Air (MBA) laptops with excitement. It has been a long time since Apple had updated the design of the MB and while the MBA V1 was nice but it had a number of public pitfalls. Needless to say Apple released some sweet new hardware that really pushed the design envelope and the lust factor of most users & developers.

After being bated with the new laptops I began asking around to see if anyone was interested in purchasing my 1.5 year old 17” MBP. Believe it or not I actually found someone quite quickly. They were interested in the 17” as a mobile video editing platform. The price was right and the timing was flexible, what more could I ask for?

I then began my usual death spiral of paralysis of analysis (as my friends kindly call it) where I overanalyze the new machines and my usage patterns. I went back and forth between the MBP, MB, and MBA. My primary requirements for a new machine are:

1. It had to be lighter than the 17”, commuting daily with that thing was becoming a chore.
2. My iPhone application development on Fairway (AppStore) and a yet to be released iPhone application couldn’t be negatively affected.
3. I didn’t want the overall quality of my mobile experience to go down. What I mean by this is the quality of the screen and overall hardware.

After analyzing the different machines for a week and half I had decided on the MB. It had mobile prowess and still appeared to have the development power and configurability I was looking for out of a laptop. It wasn’t until I actually used the machine at the Apple store that I actually decided on the MB. No I didn’t purchase it at that point, I had to wait until my buyer was ready for a swap. Needless to say I contacted him and setup a swap date after I decided.

Then a flaw in my decision was found, earlier this week I found this thread on MacRumors. The thread brought about a rather lengthy debate with a lot of detail about quality issues people have witnessed in their new MB laptop LCDs. While I was comparing machines at the Apple store my biggest complaint was that I was never able to actually see any of the machines side by side. I always had to walk back and forth which was distracting and really prevented a solid comparison. Anyhow, people were seeing huge differences in the quality of their displays and despite editing of the profiles the displays weren’t quite up to snuff with those in the MBA or MBP. One of my requirements of a new laptop was to not degrade my laptop experience and after seeing some of the screenshots I couldn’t chance it with the MB.

Fast forward (because this is getting kinda long) to Thursday evening I was awaiting peoples arrival of their new MBA. No one had actually seen one yet let alone actually ran any tests to compare their performance. I know I would have been happy with the new 15” MBP but I really wanted something lighter than the 17”, it was only 1 pound lighter.

On friday morning a MBA V2 early impression thread showed up on MacRumors along with a review … I was hooked. The specs looked awesome and the display was stunning. I began calling around and found that the Schaumburg IL Apple Store actually had the 128GB SSD in stock. I hadn’t decided on HD or SSD yet but them having it in stock pretty much decided for me. I had 3 days to buy a new machine and migrate all my crap to it in order to not lose the sale of my 17”.

Ok enough with the backstory already right? On to the review data. Remember, I’m migrating from:

17” MacBook Pro (Mid 2007)

2.4 GHz Core 2 Duo, 4MB L2, 4GB RAM, 800 MHz Bus, GeForce 8600M GT w/256 MB VRAM

to the new:

13.3” MacBook Air (V2.0 Late 2008)

1.86 GHz Core 2 Duo, 6MB L2, 2GB RAM, 1.07 GHz bus, GeForce 9400M Video w/256 MB VRAM

To start I did the usual Xbench comparison. Below are the numbers:

• MacBook Pro score 101.76
• MacBook Air score 131.98

You can also see them in a side-by-side comparison. What first struck me was that the MBA scored better than a beefier MBP, even though it was older it was unexpected. What you then notice (I knew this before hand before buying it as well) is that the performance of the MBA’s SSD skews the numbers a bit. Still surprising though is that the CPU in the MBA is no slouch beating the MBP in threading, computation, and a number other tests. That combined with a very solid graphics card in the new GeForce 9400M and this machine is “on paper” a knockout. What about actual usage?

Enter Xcode, my primary mobile computing application. While I use my computer for a number of things I spend most of my productive time writing code in Xcode working on Fairway and other applications. To perform this test I figured I would first compare my iPhone application and then a beefier mac application, Adium.

First Fairway @ ~28,000 lines of code. I’m compiling this iPhone application into Debug mode for the Device (which includes compression and signing):

MacBook Pro

19 seconds (fresh Xcode start)

13 seconds (clean all and then recompile)

MacBook Air

15 seconds (fresh Xcode start)

15 seconds (clean all and then recompile)

Not too shabby … the MBA performance is consistent and generally no change when compared to the MBP which is again surprising and perfect for my needs. Now on to Adium … I did a subversion synch to the latest version on the intertubes:

svn co svn://svn.adiumx.com/adium/tags/AdiumX_1.3.2 adium-1.3.2

I then compiled using the default settings, changing nothing at all.

MacBook Pro

170 seconds (2m 50s, fresh Xcode start)

179 seconds (2m 59s, clean all and then recompile)

MacBook Air

219 seconds (3m 39s, fresh Xcode start)

219 seconds (3m 39s, clean all and then recompile)

This test really put the machines to the test causing both of their fans to kick in at about the two minute mark. It also begins to highlight the CPU differences between the MBP (2.4 GHz) and the MBA (1.86 GHz). I was surprised that they were as close as they were to be honest and expected the MBP to handily trounce the MBA by a factor of two. One thing I did notice while running these tests was that the MBA “seemed” to perform the copy phases of the build faster than MBP, again the SSD comes to the rescue. I think this helped it keep up with the MBP as the Adium project has a lot of resources being copied around to build that awesome chat application.

Now on to a few fun tests. A number of people had seen issues with MBA V1 dropping cores and frames while running many videos and wanted to know how V2 would behave. The common test apparently was to run 3 YouTube videos concurrently and the check out iStat Pro and then watch the cores, temperature, and fans. Here are the results from running this test:

Temperature

72-78 degrees

Exhaust Fan RPM

2945 – 6040 RPM (beginning to end of 3 videos)

CPU

User 40, System 11, Nice 0, Idle 49 (during tests)

I ran the test on 3 random videos that were around 2 minutes each. I didn’t notice any cores drop out and the only odd behavior was that the fans kicked in after 2 of the videos finished and the final one was still playing. I think it was just timing as the RPMs were slowly climbing on the exhaust fan.

Finally I wanted to see how the MBA would hold up while playing WoW (World of Warcraft). I know, this isn’t a gaming machine but that doesn’t mean I don’t want to play an occasional game. These test were performed in Orgrimmar last night during a lot of Halloween festivities so a lot of effects were being rendered onscreen.

WoW Max Settings 1280×800 (Shadows Off, video)

79 degrees, Fan @ 6200 RPM, Video 17-27 FPS

WoW Max Settings 1024×768 (Shadows Off, video)

78 degrees, Fan @ 6200 RPM, Video 22-29 FPS

WoW Medium Settings 1280×800 (Shadows Off, video)

77 degrees, Fan @ 6200 RPM, Video 18-29 FPS

Again, they were much better than I expected. I could actually easily play WoW with the MBA!! I ran the tests for about 30 minutes tweaking the settings repeatedly. The medium test gave the best overall experience in game. This is especially true when you go to lesser populated areas, the FPS shoots up even more.

Finally some random tidbits of data and commentary:

• Startup of my MBP takes 95 seconds while startup of my MBA w/SSD takes 25 seconds.
• Shutdown of my MBP takes 10 seconds while shutdown of my MBA takes 3 seconds.
• The MBA drives a 24” cinema display quite well. It never froze or crashed (some people reported problems with V1 of the MBA)
• I migrated all of my data having the USB ethernet dongle connected to a powered USB hub along with keyboard, mouse, and USB HD all connected to the MBA single USB port. It was smooth and the overall throughput was good. I’ll try to do some USB networking tests and report back the results.
• The angle of the MBA keyboard is much easier on the wrist than that of the MBP. Makes for writing blog entries and code easier on the body.
• Remote disk works great mounting the DVD drive of a new model iMac (my wife’s). For reference I have a gigabit network in my home.
• The backlighting of the MBA keyboard is far superior to that of the older MBP. It’s attractive on the black keys and highly usable.
• I ran PiXel Check and I have no dead or stuck pixels, a pristine display.

In summary I really enjoy the new MBA and recommend that people really consider it when purchasing their next laptop. Apple seems to have addressed a lot of the issues with the first MBA release. For me it’s quite a change from the screen size of the 17” MBP but once I was past the initial shock the rest of the experience was flawless. The MBA addresses all of my requirements for a new mobile laptop. I’m able to continue development on Fairway and other applications, commute with less weight, and the design / hardware is top notch. The screen has amazing color ranges, I can still play games, videos are smooth, and damn it’s a nice looking machine. All in all a great machine that I highly recommend.

Note – If there’s something you’d like me to test just comment below and I’ll see what I can do.

Update 1

Some people asked that I run MSR Tools. It’s a tool that lets you see the Actual CPU usage of your cores on the MBA. It really allows you to see some of the problems like core drop or CPU throttling that were occurring in V1 of the MBA.

After running it with a the “/dev/null” test (two Terminal windows running yes > /dev/null) I saw the following:

CPU 1

800 MHz – 1.86 GHz

CPU 2

800 MHz – 1.86 GHz

After running the first null test the first CPU pegs to 1.86 GHz. Upon launching the second window the CPU begins to fluctuate between 800 MHz and 1.86 GHz. This confirms what NC Macguy is seeing as well. He also ran another test with 3 videos playing and was able to get the CPU to stay at or near 1.86 GHz. I have to set that test up still but since I’m working on cycling my battery it may have to wait until later.

I don’t think the numbers above are shocking. Yes it’s weird that the CPU randomly drops down to 800 Mhz running the null test. The computer is however being presented with conflicting problems … a long running set of CPU tasks that are raising the temperature and a need to keep the machine running which requires lowering the temperature. Because it can’t offload any of the load to the GPU (like in the video test can) it throttles down the CPU and it begins to bounce around to find some equilibrium. The machine never locks up like it seemed to do in MBA V1 so that’s a big win. All in all I think the machine handled the test properly.