First benchmarks: MacBook Air is the slowest Apple machine on the block
By Ryan Block 
posted January 24th 2008 5:43PM
posted January 24th 2008 5:43PM
- All machines tested with Xbench 1.3. Sorry, we didn't have a Mac Pro around.
- All machines tested were using Leopard (except the MBP in the center). Also, MacBook uses 1GB RAM.
- You can check the bold Xbench scores to compare the cumulative results for each test.
| Air (1.6GHz Core 2 Duo) | MBP (2.16GHz Core Duo) | MBP (2.2GHz Core 2 Duo, Tiger) | MacBook (2.2GHz Core 2 Duo) | iMac (2.4GHz Core 2 Duo) | |
|---|---|---|---|---|---|
| CPU | 79.98 | 90.06 | 112.93 | 126.66 | 138.58 |
| GCD Loop | 9.67 Mops/s | 14.89 Mops/s | 13.17 Mops/s | 13.43 Mops/s | 14.91 Mops/s |
| Floating Point Basic | 2.03 Gflop/s | 2.34 Gflop/s | 2.96 Gflop/s | 2.95 Gflop/s | 3.23 Gflop/s |
| vecLib FFT | 1.71 Gflop/s | 1.68 Gflop/s | 2.98 Gflop/s | 3.36 Gflop/s | 3.66 Gflop/s |
| Floating Point Library | 12.82 Mops/s | 15.78 Mops/s | 14.15 Mops/s | 17.80 Mops/s | 19.43 Mops/s |
| Thread Test | 148.81 | 185.75 | 219.18 | 186.4 | 208.77 |
| Computation | 2.77 Mops/s | 3.36 Mops/s | 4.04 Mops/s | 3.58 Mops/s | 3.56 Mops/s |
| Lock Contention | 7.04 Mlocks/s | 9.07 Mlocks/s | 10.47 Mlocks/s | 8.48 Mlocks/s | 11.06 Mlocks/s |
| Air (1.6GHz Core 2 Duo) | MBP (2.16GHz Core Duo) | MBP (2.2GHz Core 2 Duo, Tiger) | MacBook (2.2GHz Core 2 Duo) | iMac (2.4GHz) | |
| Memory Test | 140.42 | 129.56 | 137.1 | 150.23 | 150.82 |
| System | 143.51 | 139.25 | 126.92 | 158.95 | 151.56 |
| Allocate | 718.86 Kalloc/s | 731.25 Kalloc/s | 401.22 Kalloc/s | 856.78 Kalloc/s | 657.80 Kalloc/s |
| Fill | 5770.30 MB/s | 5565.31 MB/s | 6490.47 MB/s | 6480.99 MB/s | 6606.88 MB/s |
| Copy | 2802.78 MB/s | 2652.96 MB/s | 2954.03 MB/s | 2914.92 MB/s | 3014.12 MB/s |
| Stream | 137.46 | 121.14 | 149.05 | 142.41 | 150.08 |
| Copy | 2621.64 MB/s | 2400.99 MB/s | 2923.94 MB/s | 2799.64 MB/s | 2926.68 MB/s |
| Scale | 2602.03 MB/s | 2412.90 MB/s | 2918.16 MB/s | 2797.66 MB/s | 3022.24 MB/s |
| Add | 3230.58 MB/s | 2703.93 MB/s | 3359.19 MB/s | 3196.17 MB/s | 3364.41 MB/s |
| Triad | 3199.37 MB/s | 2681.42 MB/s | 3368.41 MB/s | 3211.97 MB/s | 3328.48 MB/s |
| Air (1.6GHz Core 2 Duo) | MBP (2.16GHz Core Duo) | MBP (2.2GHz Core 2 Duo, Tiger) | MacBook (2.2GHz Core 2 Duo) | iMac (2.4GHz) | |
| Quartz Graphics Test | 96.89 | 144.23 | 141.5 | 154.32 | 193.4 |
| Line | 6.94 Klines/s | 8.84 Klines/s | 9.23 Klines/s | 9.69 Klines/s | 11.64 Klines/s |
| Rectangle | 32.23 Krects/s | 49.67 Krects/s | 51.59 Krects/s | 51.66 Krects/s | 70.02 Krects/s |
| Circle | 7.22 Kcircles/s | 11.31 Kcircles/s | 13.30 Kcircles/s | 11.54 Kcircles/s | 15.29 Kcircles/s |
| Bezier | 2.49 Kbeziers/s | 3.58 Kbeziers/s | 3.71 Kbeziers/s | 3.79 Kbeziers/s | 4.51 Kbeziers/s |
| Text | 5.53 Kchars/s | 9.12 Kchars/s | 6.65 Kchars/s | 10.39 Kchars/s | 12.66 Kchars/s |
| OpenGL Graphics Test | 17.26 | 175.06 | 129.88 | 23.36 | 152.66 |
| Spinning Squares | 21.89 frames/s | 222.08 frames/s | 164.76 frames/s | 29.64 frames/s | 193.65 frames/s |
| User Interface Test | 105.81 | 217.79 | 303.98 | 244.28 | 335.18 |
| Elements | 485.60 refresh/s | 999.54 refresh/s | 1.40 Krefresh/s | 1.12 Krefresh/s | 1.54 Krefresh/s |
| Air (1.6GHz Core 2 Duo) | MBP (2.16GHz Core Duo) | MBP (2.2GHz Core 2 Duo, Tiger) | MacBook (2.2GHz Core 2 Duo) | iMac (2.4GHz) | |
| Disk Test | 24.05 | 27.59 | 38.13 | 39.64 | 80.72 |
| Sequential | 42.21 | 49.22 | 59.81 | 66.07 | 115.15 |
| Uncached Write | 30.96 MB/s [4K blocks] | 31.32 MB/s [4K blocks] | 42.60 MB/s [4K blocks] | 53.34 MB/s [4K blocks] | 72.17 MB/s [4K blocks] |
| Uncached Write | 31.19 MB/s [256K blocks] | 28.37 MB/s [256K blocks] | 39.19 MB/s [256K blocks] | 47.63 MB/s [256K blocks] | 66.51 MB/s [256K blocks] |
| Uncached Read | 7.27 MB/s [4K blocks] | 12.11 MB/s [4K blocks] | 11.59 MB/s [4K blocks] | 10.83 MB/s [4K blocks] | 27.81 MB/s [4K blocks] |
| Uncached Read | 30.42 MB/s [256K blocks] | 28.65 MB/s [256K blocks] | 39.37 MB/s [256K blocks] | 49.62 MB/s [256K blocks] | 69.83 MB/s [256K blocks] |
| Random | 16.81 | 19.16 | 27.99 | 28.31 | 62.13 |
| Uncached Write | 0.57 MB/s [4K blocks] | 0.65 MB/s [4K blocks] | 1.08 MB/s [4K blocks] | 1.03 MB/s [4K blocks] | 2.67 MB/s [4K blocks] |
| Uncached Write | 18.35 MB/s [256K blocks] | 17.49 MB/s [256K blocks] | 19.24 MB/s [256K blocks] | 22.73 MB/s [256K blocks] | 48.45 MB/s [256K blocks] |
| Uncached Read | 0.35 MB/s [4K blocks] | 0.46 MB/s [4K blocks] | 0.41 MB/s [4K blocks] | 0.48 MB/s [4K blocks] | 0.63 MB/s [4K blocks] |
| Uncached Read | 13.28 MB/s [256K blocks] | 15.08 MB/s [256K blocks] | 16.33 MB/s [256K blocks] | 19.31 MB/s [256K blocks] | 27.08 MB/s [256K blocks] |
But hey, uh it's thin. Doesn't that make up for every one of its disadvantages?
Maybe everything but the price tag
why the hell did you take that seriously
lol
oh, and its sexy. again, that must make up for all the shortcomings, right?
MacBook Air is Palm's Foleo in Apple disguise.....
.....and Engadget seems to be having lots of fun today driving nails into MacBookAir's coffin...
Anyone reading Engadget today will probably cancel their plans to buy one...
well, look at the bright side
its the one day engadget goes all anti-apple fanboy on apple
Wow... someone went to all that trouble to figure out that a 1.6Ghz Core 2 Duo is slower than 2Ghz+ Core 2 Duos processors. How about battery life tests?
Thats a STUPID test.. the slowest compared mac had a 500mhz faster processor.. its ~25% faster to begin with!
@Satish
Too good to be true. People are dumber than you might think.
I reckon you forgot to benchmark the Mac Mini and tell us whether or not it was faster than the Air.
@ vcx & David
The clock isn't what's important. Even the MIPS is that important. Only performance comparison yields accurate results of power and speed.
The 2.66 GHz Pentium D 805 is half as fast as the 2.4 GHz C2D E6600. It's because of other factors, such as FSB, CPU Cache, and memory timings. http://www.bcchardware.com/index.php?option=content&task=view&id=2869&Itemid=40&limit=1&limitstart=2
Yes, all the processors in Mac's test are C2D, but the sockets, motherboards, and RAM are all different.
I don't mean to make you look like n00bz, but you aren't helping.
"Even the MIPS is that important."
Even the MIPS isn't that important.
well, it's not supposed to be a blazing fast machine. i am not getting one, but if i was traveling, doing lots of presentation, i probably would. it's a marketing machine for business users. i am sticking with plans to purchase a macbook pro after its next upgrade, when gestures are added to the trackpad.
Hey, it's ultra thin, but I hope Ryan would do some drop tests and tell us if the MBA would still work after a drop from waist high, chest high, and about eye-level high. For such a device, speed does not matter, but durability.
Wow, the reasons to not buy this just keep coming.
it's an ultra portable?
13 inches are NOT Ultra portable
Ultra thin ≠ Ultra Portable.
I think it wouldn't have matter if the thing had a 7"~10" screen on it with the same small motherboard and a form factor around the size of an EEE or Cloudbook. IMHO!
...and cost around 1300 dollars less
Cloudbook rules! In fact I request a "shoot out" between the MacBook Air, the Eee PC and my beloved CloudBook. Conducted by Ryan Block, hyper nerdy performance stats, snotty comments, et al.
Agreed. Apple's lost a great opportunity to jump into the ultraportable bandwagon. I myself am a PC junkie, but my cousin, who's a real Mac-head, was waiting anxiously for the 'ultraportable Macbook El-Stevo's gonna present' on Macworld to replace his 12" G4. He's one disappointed fellow, I can tell you, and so are many other Mac-evangelistas out there.
JAmerican: This would have been what I personally expected from this year's Macworld. If it had tablet style functionality with some if hhe iPhone's interface elements, I would have preordered one by now...
But maybe it's just me, .00001% of the market...
Actually, ultra-thin does equal ultra portable. Imagine the vast majority of users for this... they'll have their MBA in their briefcases with several documents, spreadsheets, project plans, etc. all printed out on A4 (or Legal, for our imperial cousins!). That's the same footprint as an MBA (give or take -- both will fit in a manilla envelope, anyway).
The day all the documents I need to deal with are in A5/whatever the half-legal size is, will be the day a consider the smaller footprint of an Eee PC to be important.
@ Richy
Briefcases are for geezers. I can't imagine a geezer buying a MBA.
Thanks Engadget!
Note to self: slower processor = slower computer.
I think I speak for all of us when I say that that needed clearing up.
The 1.8" drive + graphics matter too you know.
There's more to the MBA's poor performance than just the slower processor.
I was interested to see just how horrible the hard drive is, but I'd have to say that the real surprise here is how the MBA scored less than half of what the $500 cheaper MacBook scored on the "User Interface" test.
Obviously, the processor doesn't help, but it's not just that. If it was just a slower processor, that might be OK for some people. Unfortunately, as these tests show, everything in the MBA sucks much more than the much cheaper MacBook.
Drive access speed is a huge factor in Mac OS X's performance. Testing with the 80 GB iPod drive is going to be hellishly slow. I would much prefer a 32 GB SSD in there for a $300 bump than the 80 1.8", or the $1000 64 GB SSD.
Am I the only one who was disappointed to see my $2k MBP beaten by a $1k Macbook in everything but Open-GL? Even when the processors were the same...
Did we seriously not expect this?
Yeah, I mean I think it would be strange if it outperformed other computers. I can definitely sense that the reason most people love to hate this product isn't because it's slow... Or because network adapters don't fit without a short extension cable... It's because it's EXPENSIVE for all the opportunity cost that comes with it. And, though I'm an Apple fan and stockholder, that's definitely understandable.
Why not benchmark against other ultraportables using BootCamp so we can get an idea as to how it stacks up against the competition?
The Air is an ultra-portable?
Wow. Buggy comment system. It put my first reply in the wrong place. To repeat and make it clear who I'm replying to:
The Air is an ultra-portable?
Because it's not an ultraportable.
Twice. Great. Those are supposed to be replies to pop. (So is this one.) Is there a way to file a bug report for the comment system?
Because it's not an ultraportable.
Now, you'd have to take into consideration that loading Windows using BootCamp would not yield the most accurate number because Win is not running natively. I'm assuming here. I've never used BootCamp, so I don't know the particulars of it.
No contest. The base XPS M1330 would kill this piece of junk. If you bump up the M1330 to the price of the MBA, it's performance would be on par with the MacBook Pro (2.2GHz, 4GB RAM, 320GB HD, 128MB GeForce).
Wireless broadband is an option on the 1330. Also, the 64GB SSD is only a $675 option on the Dell. Sure it's a 2.5" instead of 1.8", but whatever.
I really don't understand why anyone would buy a MacBook Air. Even the richest Mac fanboy in the world should be able to see that it's just a bad deal.
Because everyone else messed this up:
Its not an ultraportable.
Oh right! Because it ISN'T an ultraportable. It's a thin laptop. Yes, there's a difference. My eee fits places better than the Air would, even if it is 3 times as thick. Oh, it also weighs less, boots faster, and costs less than a quarter of the low end air. hmm. The air had better win with that sort of price difference.
Testing the broken comment system to say that it's not an ultraportable.
Doesn't matter - you can still slice cheese with it. That's what it's for, right? Wealthy yuppies who go to Napa Valley on the weekend?
Woah, please blogsmith, fix the comment system.
Rob - bootcamp runs windows natively. It's basically a set of native windows drivers for the mac hardware, plus a way to boot windows -- OSX never loads. Parallels is a hosted virtual environment, and would be the one you would not want to use as a comparison.
Why would you make a slim notebook thats not REALLY an ultra-portable (given the 13.3" screen, full keyboard, etc), price it up the wazoo, only to have it run dog shit slow??
Because once you make it pretty and slap the Apple logo on it, people will buy it.
dog feces has a speed?
yes it does, and MBA is slower than that