My new 13" Macbook Air vs 2011 13" Pro

I've only been using my new 13" Macbook Air (base i5 with 4GB and 128GB SSD) since last night, but I'm already majorly impressed. Just a few random things I can think of, will add more as I discover other things.

-The screen, 1440x900 vs 1280x800, it is so much more spacious. I can now put the dock on the left and use most of the 900 vertical pixels for actual work. It actually seems brighter than the MBP screen. Text and everything looks much more crisp. This was the major selling point for me.

-Even though the screen is still semi-reflective, it is a LOT less reflective than the MBP - I can no longer see a clear reflection of myself.

-The speakers, I can't believe how loud and dare I say full-sounding this thing is. Probably because Apple managed to fit the speakers into an enclosure of sort, where as in the 13" MBP it's just a speaker sitting in there. You can see this "enclosure" from the iFixit teardown pics.

-Battery life *seems* to be better, even though they're both rated for 7hrs by Apple. When I first opened it, it registered 91% full. I've been messing with it for a good 4-5 hours combined now and it's still reading 18%.

-The only slight con for me is the keyboard. I prefer the longer key travel of the MBP. But this is to be expected on a laptop so thin. Oh the F4 key, on the MBP even running Lion toggles the Dashboard, which I prefer (I'll never used the Launchpad). Yes on the Air it now triggers Launchpad. There seems to be a bug on Lion where if you clicked "restore to defaults" in the keyboard shortcut settings in system preferences, under "launchpad & dock", the "show launchpad" gets unchecked. If you try to check it again, it will insist that you pick a custom function key (such as fn-F4). If you don't assign it a custom key you can no longer check this option. This happens on both my MBA and the MBP running Lion. On the MBA I now lost that function key. I created a new user ID and it worked fine until I clicked restore to defaults again. On the MBP it continues to trigger the Dashboard. So if you're using the new MBA do not click that restore to default button or you'll lose the F4 function. If anyone knows how to restore it please let me know.

-Performance, saved it for last It is every bit as quick as the 13" MBP which has a 64GB Sandforce SSD installed. I ran Xbench on both and the results are very close. I had the last gen 11.6" MBA for a little while, on some things (namely heavy Flash sites) it was unbearably slow so I let it go. Everything I tossed at this new MBA so far is every bit as quick as the MBP. I don't doubt that if you run some very CPU intensive tasks on it the full 35W CPU's will outperform it, but I only do those tasks on my desktops.

I've been a Windows guy until just a couple years ago, professionally and personally, now I use my iPhone/ipad/macbook more than my Windows computers. There are just so many little things that you go "wow" or "cool" in OSX, now more so in Lion, in Windows mostly it's like "hmm OK". I can see a typical person after seeing a Macbook and a Windows laptop will likely pick the Macbook, if they have the budget to do so.

Here's the Xbench 1.3 result. Xbench can't perform the thread test in Lion so that was the only thing that was unchecked.

2011 13" Macbook Air (base i5 with 4GB RAM and 128GB SSD):

Results 199.90
System Info
Xbench Version 1.3
System Version 10.7 (11A2063)
Physical RAM 4096 MB
Model MacBookAir4,2
Drive Type APPLE SSD SM128C
CPU Test 195.62
GCD Loop 261.68 13.79 Mops/sec
Floating Point Basic 171.29 4.07 Gflop/sec
vecLib FFT 120.37 3.97 Gflop/sec
Floating Point Library 403.08 70.19 Mops/sec
Memory Test 509.70
System 506.84
Allocate 1934.24 7.10 Malloc/sec
Fill 351.28 17080.24 MB/sec
Copy 391.35 8083.13 MB/sec
Stream 512.60
Copy 503.00 10389.17 MB/sec
Scale 496.30 10253.46 MB/sec
Add 532.93 11352.58 MB/sec
Triad 519.75 11118.72 MB/sec
Quartz Graphics Test 314.77
Line 298.78 19.89 Klines/sec [50% alpha]
Rectangle 394.51 117.78 Krects/sec [50% alpha]
Circle 299.95 24.45 Kcircles/sec [50% alpha]
Bezier 269.92 6.81 Kbeziers/sec [50% alpha]
Text 337.34 21.10 Kchars/sec
OpenGL Graphics Test 127.85
Spinning Squares 127.85 162.18 frames/sec
User Interface Test 109.34
Elements 109.34 501.81 refresh/sec
Disk Test 357.54
Sequential 226.85
Uncached Write 435.32 267.28 MB/sec [4K blocks]
Uncached Write 335.09 189.60 MB/sec [256K blocks]
Uncached Read 99.98 29.26 MB/sec [4K blocks]
Uncached Read 425.70 213.96 MB/sec [256K blocks]
Random 843.43
Uncached Write 789.93 83.62 MB/sec [4K blocks]
Uncached Write 623.51 199.61 MB/sec [256K blocks]
Uncached Read 1822.11 12.91 MB/sec [4K blocks]
Uncached Read 755.31 140.15 MB/sec [256K blocks]

2011 13" Macbook Pro (base i5 with Sandforce SSD and 6GB RAM):
Results 216.58
System Info
Xbench Version 1.3
System Version 10.7 (11A511)
Physical RAM 6144 MB
Model MacBookPro8,1
Drive Type SSD G2 series 64GB
CPU Test 220.86
GCD Loop 290.26 15.30 Mops/sec
Floating Point Basic 193.05 4.59 Gflop/sec
vecLib FFT 139.30 4.60 Gflop/sec
Floating Point Library 433.37 75.46 Mops/sec
Memory Test 415.52
System 517.02
Allocate 2128.02 7.81 Malloc/sec
Fill 327.96 15946.13 MB/sec
Copy 437.95 9045.58 MB/sec
Stream 347.33
Copy 337.25 6965.84 MB/sec
Scale 332.46 6868.52 MB/sec
Add 363.85 7750.78 MB/sec
Triad 357.80 7654.13 MB/sec
Quartz Graphics Test 358.30
Line 335.70 22.35 Klines/sec [50% alpha]
Rectangle 463.86 138.49 Krects/sec [50% alpha]
Circle 348.78 28.43 Kcircles/sec [50% alpha]
Bezier 302.31 7.62 Kbeziers/sec [50% alpha]
Text 378.06 23.65 Kchars/sec
OpenGL Graphics Test 146.92
Spinning Squares 146.92 186.38 frames/sec
User Interface Test 132.47
Elements 132.47 607.96 refresh/sec
Disk Test 276.02
Sequential 165.56
Uncached Write 223.87 137.45 MB/sec [4K blocks]
Uncached Write 198.84 112.50 MB/sec [256K blocks]
Uncached Read 84.21 24.64 MB/sec [4K blocks]
Uncached Read 358.55 180.21 MB/sec [256K blocks]
Random 829.28
Uncached Write 1384.96 146.61 MB/sec [4K blocks]
Uncached Write 393.34 125.92 MB/sec [256K blocks]
Uncached Read 2786.35 19.75 MB/sec [4K blocks]
Uncached Read 833.21 154.61 MB/sec [256K blocks]

 

Nice, thanks! Interesting that the GPU in the MBP scores a tad higher despite being the same, the GPU also has Turbo Boost when the TDP of the processor allows it, with less thermal dissipation maybe the GPU gets to boost less in the Air.

There was a rumor that the SSD in these would be faster than the Core 2 MBA, is that true? Have you tried running any storage specific benchmarks?

 

GPU probably scored higher in the MBP13 because it has a larger chassis, lending to better cooling of the gpu, and therefore it can run at 100% for longer periods of time.

And that rumor depends on the speed of the SSD. However, a MBP13 with SSD would be sick nasty

 

Neither one of these laptops has a GPU - both have an Intel HD 3000 IGP (Integrated Graphics Processor). Macbook Pro has a full voltage HD 3000 that can clock form 650-1200 MHz (base - turbo boost). Macbook Air has an ultra low voltage HD 3000 that is underclocked to 350-1000 MHz (base - turbo boost).

 

Ah, that would explain it. I did not know the ULV ones were clocked differently. I wonder how often they spend in the Turbo state then? The GPU (IGP's are still called that, you can differentiate with discreet or integrated, ie dGPU) scored differently but pretty close, they must be in their turbo frequencies often as the standard clock would yield much lower scores in the low voltage one.

 

Also, Sandforce drives are gas guzzlers in the SSD world than Apple's presumed Toshiba drives, and obviously the ULV eats up less power. Furthermore, there's the lack of optical drive. I'd be surprised if it didn't have a longer battery life.

 

While the resolution on the MBA is higher, the gamut and viewing angles on the MBP are better (77% vs. 47% according to Anandtech).

That's not true. MBA 2010 and 2011 both use the same Samsung and Toshiba drives. The Samsung being slightly faster.

 

Small correction, its 350-1200 and 650-1300. Since 350 is further from 650 than 1200-1300 and the scores are pretty close, it seems like they spend most of the time in Turbo speeds, from this benchmark.

Which begs the question, will either one (esp the Air) slow down as the thermal limits are reached after a while of gameplay/graphics work? Because it only boosts when the thermals allow it.

Intel® Core? i5-2557M Processor (3M Cache, 1.70 GHz))

Intel® Core? i5-2415M Processor (3M Cache, 2.30 GHz)

 

That is the big question that I haven't seen any answer too.

And how quickly will they slow down when they're running on turbo boost.

And how much difference will there be in thermal throttling between 11" and 13" Air.