11-inch MacBook Air Xbench Results

[ChaosAngel]

Hi all,

I popped in to my local iStore today (unofficial Apple Store) and to my surprise they had both the 11-inch and 13-inch MacBook Air's on demo.

I only had a few mins, but took the opportunity to install and run some Xbench tests on the 11-inch model.

I have posted the results below:

http://www.lifeintech.co.uk/gallery/11-inch-macbook-air-xbench-results/

I have to say, before I saw the two models in reality I was convinced that the 13-inch would be the model for me. However after comparing them next to each other and using the 11-inch model I honestly believe it is the better product. The screen although small, is beautiful. Literally a millions times better then any NetBook I have seen.

Thankfully they also had the 11-inch hocked up to a 24" Cinema Display and I can confirm that it also worked amazingly well. I opened multiple applications, played some video (Quicktime 720p) and surfed the web. No noticeable slow down and a perfectly usable experience.

Hope this helps some people with there buying decision.

 

[jclardy]

Well if I buy one I am really looking forward to the SSD. In some of the random tests it is around 10x faster than the 5400 RPM in my Mac Mini.

Although the Open GL test was a little disappointing as my Mac Mini with 9400M scored higher than the 320M which is a bit strange, I guess it could also be affected by the processor speed.

But I am definitely still considering the 11 inch one (and keeping my 15" i7 as a more mobile desktop machine).

 

[ChaosAngel]

If I get the chance tomorrow I'll see if I can post some more results. I will also post the results for my current MacBook Pro (Rev G) for comparison.

I am seriously considering the 11" MacBook Air.

 

[J400uk]

According to this review http://www.chipchick.com/2010/10/macbook-air-11-inch-review.html it should score about 125 in xbench overall.

 

[AtmChm]

Hi all,

I popped in to my local iStore today (unofficial Apple Store) and to my surprise they had both the 11-inch and 13-inch MacBook Air's on demo.

I only had a few mins, but took the opportunity to install and run some Xbench tests on the 11-inch model.

I have posted the results below:

http://www.lifeintech.co.uk/gallery/11-inch-macbook-air-xbench-results/

Interesting. It appears the 11" MBA is slower than my late Rev C with TS128B SSD in all categories except the disk test, where it is marginally better (220 vs. 210). I'm not surprised the CPU is slower, but the enhanced graphics are clearly slower. Here are the xbench results for my machine. I'd really like to see results posted for the 13" model.


Results 173.61
System Info
Xbench Version 1.3
System Version 10.6.4 (10F569)
Physical RAM 2048 MB
Model MacBookAir2,1
Drive Type APPLE SSD TS128B
CPU Test 151.43
GCD Loop 250.03 13.18 Mops/sec
Floating Point Basic 121.67 2.89 Gflop/sec
vecLib FFT 100.11 3.30 Gflop/sec
Floating Point Library 237.66 41.38 Mops/sec
Thread Test 222.82
Computation 272.76 5.53 Mops/sec, 4 threads
Lock Contention 188.33 8.10 Mlocks/sec, 4 threads
Memory Test 178.09
System 208.42
Allocate 274.92 1.01 Malloc/sec
Fill 169.83 8257.51 MB/sec
Copy 205.41 4242.58 MB/sec
Stream 155.46
Copy 149.43 3086.46 MB/sec
Scale 146.66 3030.02 MB/sec
Add 163.82 3489.68 MB/sec
Triad 163.53 3498.29 MB/sec
Quartz Graphics Test 155.72
Line 138.70 9.23 Klines/sec [50% alpha]
Rectangle 171.82 51.30 Krects/sec [50% alpha]
Circle 147.38 12.01 Kcircles/sec [50% alpha]
Bezier 164.10 4.14 Kbeziers/sec [50% alpha]
Text 161.27 10.09 Kchars/sec
OpenGL Graphics Test 125.63
Spinning Squares 125.63 159.36 frames/sec
User Interface Test 222.38
Elements 222.38 1.02 Krefresh/sec
Disk Test 211.24
Sequential 140.39
Uncached Write 191.75 117.73 MB/sec [4K blocks]
Uncached Write 191.10 108.12 MB/sec [256K blocks]
Uncached Read 66.43 19.44 MB/sec [4K blocks]
Uncached Read 334.28 168.01 MB/sec [256K blocks]
Random 426.53
Uncached Write 262.94 27.83 MB/sec [4K blocks]
Uncached Write 315.77 101.09 MB/sec [256K blocks]
Uncached Read 1249.48 8.85 MB/sec [4K blocks]
Uncached Read 622.02 115.42 MB/sec [256K blocks]

Update: a quick check of the xbench posting site shows the 13" model also with lower scores (as compared to the late Rev C MBA) in all categories except the disk test where it is about 10% higher.

 

[ChaosAngel]

I should have access to the 13-inch model today so will do my best to post the results.

 

[Dmac77]

So I see that the MBA has a score of ~99, can anyone tell me how much of a hit in performance I'll see switching to the 11.6 MBA base from a 2007 iMac 20" 2.4C2D on Snow Leopard (I got an Xbench score or 152.23). The iMac scores higher in every category except for the disk test (in which the MBA completely destroys my iMac which I expected). I know that almost no one will be able to give me a hands-on account, but theoretically how much of a hit would I see?

Will the superiority of the Flash Memory, make up for the loss in performance in other areas? I will be keeping the iMac, but there's a chance it might get put in it's box and get replaced by the MBA and a 27" Cinema Display, if the MBA will be able to handle some occasional photography work (processing RAW photos which weigh in at about 10mb, and occasional photoshop work).

-Don

 

[Hellhammer]

So I see that the MBA has a score of ~99, can anyone tell me how much of a hit in performance I'll see switching to the 11.6 MBA base from a 2007 iMac 20" 2.4C2D on Snow Leopard (I got an Xbench score or 152.23). The iMac scores higher in every category except for the disk test (in which the MBA completely destroys my iMac which I expected). I know that almost no one will be able to give me a hands-on account, but theoretically how much of a hit would I see?

Will the superiority of the Flash Memory, make up for the loss in performance in other areas? I will be keeping the iMac, but there's a chance it might get put in it's box and get replaced by the MBA and a 27" Cinema Display, if the MBA will be able to handle some occasional photography work (processing RAW photos which weigh in at about 10mb, and occasional photoshop work).

-Don

You wanted this to be theoretical so here we go

CPU: 2.4GHz vs 1.4GHz; iMac has 71.4% faster CPU (or MBA has 41.7% slower CPU, depends how you count it). iMac also has 4MB of L2 cache while MBA has 3MB. I'm just counting the GHz difference since they both use the same microarchitecture and have two cores so this is very theoretical.

RAM: 667MHz vs 1066MHz; MBA's RAM offer bandwidth of up to 68.2Gbit/s while iMac offers only 42.7Gbit/s, so MBA is 59.7% faster in memory bandwidth (note: these numbers are at single-channel, I don't know does MBA support dual-channeling).

MBA can take 4GB of RAM while your iMac can take up to 6GB, so iMac can take 50% more RAM. How much do you have now BTW?

GPU: iMac has underclocked ATI Mobility Radeon HD 2600 XT and MBA has NVIDIA GeForce 320M (not sure is it underclocked). You can do the comparison on your own, they seem to be pretty equal.

Storage: MBA's SSD is a lot faster than your 7200rpm 3.5" HD in iMac. Can't give any specific data because it depends on the HD

Due to the SSD, the MBA should feel snappier but photo editing may be a bit slower because of the slower CPU. However, it should be okay for Photoshop if you're not in hurry

 

[Dmac77]

You wanted this to be theoretical so here we go

CPU: 2.4GHz vs 1.4GHz; iMac has 71.4% faster CPU (or MBA has 41.7% slower CPU, depends how you count it). iMac also has 4MB of L2 cache while MBA has 3MB. I'm just counting the GHz difference since they both use the same microarchitecture and have two cores so this is very theoretical.

RAM: 667MHz vs 1066MHz; MBA's RAM offer bandwidth of up to 68.2Gbit/s while iMac offers only 42.7Gbit/s, so MBA is 59.7% faster in memory bandwidth (note: these numbers are at single-channel, I don't know does MBA support dual-channeling).

MBA can take 4GB of RAM while your iMac can take up to 6GB, so iMac can take 50% more RAM. How much do you have now BTW?

GPU: iMac has underclocked ATI Mobility Radeon HD 2600 XT and MBA has NVIDIA GeForce 320M (not sure is it underclocked). You can do the comparison on your own, they seem to be pretty equal.

Storage: MBA's SSD is a lot faster than your 7200rpm 3.5" HD in iMac. Can't give any specific data because it depends on the HD

Due to the SSD, the MBA should feel snappier but photo editing may be a bit slower because of the slower CPU. However, it should be okay for Photoshop if you're not in hurry

I currently have 2 GB of RAM in the in iMac (and I'm fairly certain that it came stock with 1 GB, tbh I can't remember). I am able to do just fine with the 2 GB of RAM, I'm the kind of person who closes every program when I'm done with it (a habit from my Windows days I guess). TBH, I'm not sure if the MBA supports dual-channeling and I can't find anything confirming or denying that.

A small performance hit on photo editing won't bother me, I just don't want to be beach-balling every 10 seconds.

Do you think I'll be ok?

-Don

 

[Hellhammer]

TBH, I'm not sure if the MBA supports dual-channeling and I can't find anything confirming or denying that.

It's hard to say because MBA has no RAM slots. If it had 2x1GB, then it would be easier to say. However, RAM bandwidth is VERY theoretical difference as most apps cannot take advantage of such high bandwidths so I doubt you would notice difference because of the RAM speeds.

A small performance hit on photo editing won't bother me, I just don't want to be beach-balling every 10 seconds.

In 99% of cases, beachballing occurs because of 1) page outing (i.e. lack of RAM, data has to be written to HD instead of RAM) 2) when CPU and/or other components have to wait for the HD to spin up (with SSD, you shouldn't have this problem since its access time is zillion times faster )

It shouldn't be beachballing, photo editing isn't very CPU intensive although it depends how intensive your editing is. For hobby based editing, MBA should do the job fine.

I would get 4GB RAM though, that will make it feel faster and more "future-proof" plus Photoshop loves RAM so the more you have, the faster it will be.

 

[Dmac77]

It's hard to say because MBA has no RAM slots. If it had 2x1GB, then it would be easier to say. However, RAM bandwidth is VERY theoretical difference as most apps cannot take advantage of such high bandwidths so I doubt you would notice difference because of the RAM speeds.



In 99% of cases, beachballing occurs because of 1) page outing (i.e. lack of RAM, data has to be written to HD instead of RAM) 2) when CPU and/or other components have to wait for the HD to spin up (with SSD, you shouldn't have this problem since its access time is zillion times faster )

It shouldn't be beachballing, photo editing isn't very CPU intensive although it depends how intensive your editing is. For hobby based editing, MBA should do the job fine.

I would get 4GB RAM though, that will make it feel faster and more "future-proof" plus Photoshop loves RAM so the more you have, the faster it will be.

OK, that reassures me now. I'll keep the idea of getting the 4GB of RAM in mind, but my instant gratification complex will probably get the best of me like it always does Also regarding the future-proofing aspect, that won't be much of an issue, because the MBA only has to last 1.5 years, and then I'm off to college which will inevitably result in a new MacBook of some kind.

-Don

 

[jclardy]

Is the gpu underclocked? I ran xbench on my MacBook 1st gen with GMA 950 and the OpenGL test was more than twice as high which doesn't make any sense to me...

CPU ram and the ssd were higher on the air though.

 

[Jiten]

hmm based on the test you guys got it might be probable that the 11 incher's GPU is underclocked. Perhaps for thermal reasons?

 

[Hellhammer]

hmm based on the test you guys got it might be probable that the 11 incher's GPU is underclocked. Perhaps for thermal reasons?

Previous gen's 9400M was underclocked so I would be surprised if 320M was not, especially in the 11.6".

 

[Dmac77]

I'm currently at the Briarwood Apple Store (typing this out on a 11.6" MBA) and I had a chance to run Xbench myself, and I got a score of 124.74 (1.4 GHz, 2GB RAM).

-Don

 

[AtmChm]

I'm currently at the Briarwood Apple Store (typing this out on a 11.6" MBA) and I had a chance to run Xbench myself, and I got a score of 124.74 (1.4 GHz, 2GB RAM).

-Don

OK, it's clear the 11.6 inch MBA underperforms the late Rev C MBA in all categories except the disk test. Can anyone post results from a 13.3 inch MBA? I'm expecting the 1.8 GHz stock cpu to underperform the 2.1 on the late Rev C, but it would be interesting to see the other scores.

I'm very suspicious that the new 13.3 MBA will either underperform or at best be on par with the late Rev C MBA, even with the BTO 2.1 GHz processor. This is partly a selfish suspicion since I bought an MBA in July. So far looking at the xbench postings on their site, the new MBAs are getting lower overall scores than my late Rev C MBA. However, I don't know if anyone has gotten a BTO 2.1 GHz delivered yet.

 

[biosci]

I just ran XBench on my Late 2008 MBP dual 2.8 C2D and this is what I got (See below). I also just ordered the 11.6 1.6 C2D 4GB / 128SSD.. I hope I can compare when that comes in:

Here's the MBP:

Results 145.46
System Info
Xbench Version 1.3
System Version 10.6.4 (10F569)
Physical RAM 4096 MB
Model MacBookPro5,1
Drive Type Hitachi HTS723232L9SA62
CPU Test 199.99
GCD Loop 327.04 17.24 Mops/sec
Floating Point Basic 161.27 3.83 Gflop/sec
vecLib FFT 132.37 4.37 Gflop/sec
Floating Point Library 313.65 54.62 Mops/sec
Thread Test 303.37
Computation 394.24 7.99 Mops/sec, 4 threads
Lock Contention 246.54 10.61 Mlocks/sec, 4 threads
Memory Test 196.95
System 250.17
Allocate 438.92 1.61 Malloc/sec
Fill 193.08 9388.16 MB/sec
Copy 220.54 4555.15 MB/sec
Stream 162.40
Copy 151.92 3137.83 MB/sec
Scale 159.65 3298.30 MB/sec
Add 169.55 3611.76 MB/sec
Triad 169.88 3634.19 MB/sec
Quartz Graphics Test 220.45
Line 193.78 12.90 Klines/sec [50% alpha]
Rectangle 252.14 75.28 Krects/sec [50% alpha]
Circle 213.00 17.36 Kcircles/sec [50% alpha]
Bezier 224.08 5.65 Kbeziers/sec [50% alpha]
Text 227.46 14.23 Kchars/sec
OpenGL Graphics Test 200.50
Spinning Squares 200.50 254.35 frames/sec
User Interface Test 329.83
Elements 329.83 1.51 Krefresh/sec
Disk Test 45.06
Sequential 93.44
Uncached Write 124.10 76.19 MB/sec [4K blocks]
Uncached Write 120.87 68.39 MB/sec [256K blocks]
Uncached Read 52.04 15.23 MB/sec [4K blocks]
Uncached Read 137.76 69.24 MB/sec [256K blocks]
Random 29.69
Uncached Write 9.43 1.00 MB/sec [4K blocks]
Uncached Write 102.69 32.88 MB/sec [256K blocks]
Uncached Read 85.16 0.60 MB/sec [4K blocks]
Uncached Read 138.21 25.65 MB/sec [256K blocks]

 

[hmg]

OK, it's clear the 11.6 inch MBA underperforms the late Rev C MBA in all categories except the disk test. Can anyone post results from a 13.3 inch MBA? I'm expecting the 1.8 GHz stock cpu to underperform the 2.1 on the late Rev C, but it would be interesting to see the other scores.

I'm very suspicious that the new 13.3 MBA will either underperform or at best be on par with the late Rev C MBA, even with the BTO 2.1 GHz processor. This is partly a selfish suspicion since I bought an MBA in July. So far looking at the xbench postings on their site, the new MBAs are getting lower overall scores than my late Rev C MBA. However, I don't know if anyone has gotten a BTO 2.1 GHz delivered yet.

Scores taken today in Sydney Apple Store on a demo 1.8GHz 2GB 128GB (look at the very weird disk numbers) there were no 2.13s on display:


Results 168.59
System Info
Xbench Version 1.3
System Version 10.6.4 (10F3061)
Physical RAM 2048 MB
Model MacBookAir3,2
Drive Type APPLE SSD TS128C
CPU Test 132.25
GCD Loop 218.60 11.52 Mops/sec
Floating Point Basic 105.19 2.50 Gflop/sec
vecLib FFT 87.97 2.90 Gflop/sec
Floating Point Library 208.48 36.30 Mops/sec
Thread Test 213.81
Computation 247.05 5.00 Mops/sec, 4 threads
Lock Contention 188.45 8.11 Mlocks/sec, 4 threads
Memory Test 179.99
System 197.13
Allocate 245.03 899.84 Kalloc/sec
Fill 161.23 7839.54 MB/sec
Copy 202.63 4185.26 MB/sec
Stream 165.59
Copy 158.12 3265.96 MB/sec
Scale 155.47 3211.87 MB/sec
Add 176.48 3759.39 MB/sec
Triad 174.44 3731.66 MB/sec
Quartz Graphics Test 145.30
Line 126.28 8.41 Klines/sec [50% alpha]
Rectangle 168.87 50.42 Krects/sec [50% alpha]
Circle 140.84 11.48 Kcircles/sec [50% alpha]
Bezier 146.03 3.68 Kbeziers/sec [50% alpha]
Text 151.01 9.45 Kchars/sec
OpenGL Graphics Test 120.88
Spinning Squares 120.88 153.35 frames/sec
User Interface Test 177.45
Elements 177.45 814.42 refresh/sec
Disk Test 222.62
Sequential 147.48
Uncached Write 242.11 148.65 MB/sec [4K blocks]
Uncached Write 271.39 15355 MB/sec [256K blocks]
Uncached Read 61.33 17.95 MB/sec [4K blocks]
Uncached Read 332.95 167.34 MB/sec [256K blocks]
Random 453.91
Uncached Write 279.21 29.56 MB/sec [4K blocks]
Uncached Write 399.26 127.82 MB/sec [256K blocks]
Uncached Read 1264.53 896 MB/sec [4K blocks]
Uncached Read 516.71 95.88 MB/sec [256K blocks]

 

[mulo]

jesus, code tags guys!

lalalala

 

[ChaosAngel]

Hi all,

Just posted an update for the MacBook Air Xbench results if anyone is interested:

http://www.lifeintech.co.uk/blog/2010/10/22/new-macbook-air-hands-on.html

It includes both 11-inch configurations (1.4GHz and 1.6GHz). I have also compared the results to my 15-inch MacBook Pro.

Hopefully this helps with some peoples buying decisions (especially the current MacBook Pro users).

 

[bcaslis]

I picked up the 11" ultimate configuration yesterday. If you are looking at benchmarks to decide if you want this computer or not then you don't want it. But if you want something that is just pure fun, then go get it. Seriously, despite it's puny specs it is the most fun Mac I've used in a long time. For normal tasks it just flies. But anything CPU intensive will be a task. If that is what you do, then don't bother with the 11".

Xbench scores:

Results 141.26
System Info
Xbench Version 1.3
System Version 10.6.4 (10F3061)
Physical RAM 4096 MB
Model MacBookAir3,1
Drive Type APPLE SSD TS128C
CPU Test 113.13
GCD Loop 186.93 9.85 Mops/sec
Floating Point Basic 90.43 2.15 Gflop/sec
vecLib FFT 74.71 2.46 Gflop/sec
Floating Point Library 179.71 31.29 Mops/sec
Thread Test 175.44
Computation 249.56 5.06 Mops/sec, 4 threads
Lock Contention 135.26 5.82 Mlocks/sec, 4 threads
Memory Test 149.29
System 166.95
Allocate 275.12 1.01 Malloc/sec
Fill 128.83 6264.04 MB/sec
Copy 152.15 3142.59 MB/sec
Stream 135.01
Copy 129.17 2668.03 MB/sec
Scale 126.46 2612.71 MB/sec
Add 141.89 3022.47 MB/sec
Triad 144.29 3086.79 MB/sec
Quartz Graphics Test 115.51
Line 106.68 7.10 Klines/sec [50% alpha]
Rectangle 121.60 36.31 Krects/sec [50% alpha]
Circle 104.68 8.53 Kcircles/sec [50% alpha]
Bezier 122.50 3.09 Kbeziers/sec [50% alpha]
Text 125.42 7.85 Kchars/sec
OpenGL Graphics Test 107.39
Spinning Squares 107.39 136.23 frames/sec
User Interface Test 141.37
Elements 141.37 648.80 refresh/sec
Disk Test 215.00
Sequential 157.34
Uncached Write 224.81 138.03 MB/sec [4K blocks]
Uncached Write 278.92 157.81 MB/sec [256K blocks]
Uncached Read 68.21 19.96 MB/sec [4K blocks]
Uncached Read 366.42 184.16 MB/sec [256K blocks]
Random 339.38
Uncached Write 136.42 14.44 MB/sec [4K blocks]
Uncached Write 511.37 163.71 MB/sec [256K blocks]
Uncached Read 1106.35 7.84 MB/sec [4K blocks]
Uncached Read 626.35 116.22 MB/sec [256K blocks]


Geekbench 32 bit scores:

Platform: Mac OS X x86 (32-bit)
Compiler: GCC 4.0.1 (Apple Inc. build 5494)
Operating System: Mac OS X 10.6.4 (Build 10F3061)
Model: MacBookAir3,1
Motherboard: Apple Inc. Mac-942452F5819B1C1B 1.0
Processor: Intel(R) Core(TM)2 Duo CPU U9600 @ 1.60GHz
Processor ID: GenuineIntel Family 6 Model 23 Stepping 10
Logical Processors: 2
Physical Processors: 1
Processor Frequency: 1.60 GHz
L1 Instruction Cache: 32.0 KB
L1 Data Cache: 32.0 KB
L2 Cache: 3.00 MB
L3 Cache: 0.00 B
Bus Frequency: 800 MHz
Memory: 4.00 GB
Memory Type: 1067 MHz DDR3
SIMD: 1
BIOS: Apple Inc. MBA31.88Z.0061.B00.1009101530
Processor Model: Intel(R) Core(TM)2 Duo CPU U9600 @ 1.60GHz
Processor Cores: 2

Integer (Score: 1806)
Blowfish single-threaded scalar -- 1122, , 49.3 MB/sec
Blowfish multi-threaded scalar -- 2270, , 93.0 MB/sec
Text Compress single-threaded scalar -- 1248, , 3.99 MB/sec
Text Compress multi-threaded scalar -- 2262, , 7.42 MB/sec
Text Decompress single-threaded scalar -- 1114, , 4.58 MB/sec
Text Decompress multi-threaded scalar -- 2142, , 8.54 MB/sec
Image Compress single-threaded scalar -- 1140, , 9.42 Mpixels/sec
Image Compress multi-threaded scalar -- 2092, , 17.6 Mpixels/sec
Image Decompress single-threaded scalar -- 1013, , 17.0 Mpixels/sec
Image Decompress multi-threaded scalar -- 1969, , 32.1 Mpixels/sec
Lua single-threaded scalar -- 1857, , 715.1 Knodes/sec
Lua multi-threaded scalar -- 3445, , 1.33 Mnodes/sec

Floating Point (Score: 3233)
Mandelbrot single-threaded scalar -- 1192, , 793.5 Mflops
Mandelbrot multi-threaded scalar -- 2274, , 1.49 Gflops
Dot Product single-threaded scalar -- 2186, , 1.06 Gflops
Dot Product multi-threaded scalar -- 4425, , 2.02 Gflops
Dot Product single-threaded vector -- 1747, , 2.09 Gflops
Dot Product multi-threaded vector -- 3734, , 3.88 Gflops
LU Decomposition single-threaded scalar -- 469, , 417.5 Mflops
LU Decomposition multi-threaded scalar -- 954, , 836.7 Mflops
Primality Test single-threaded scalar -- 2614, , 390.4 Mflops
Primality Test multi-threaded scalar -- 3818, , 708.7 Mflops
Sharpen Image single-threaded scalar -- 3356, , 7.83 Mpixels/sec
Sharpen Image multi-threaded scalar -- 6324, , 14.6 Mpixels/sec
Blur Image single-threaded scalar -- 4214, , 3.33 Mpixels/sec
Blur Image multi-threaded scalar -- 7965, , 6.26 Mpixels/sec

Memory (Score: 1920)
Read Sequential single-threaded scalar -- 2281, , 2.79 GB/sec
Write Sequential single-threaded scalar -- 2140, , 1.46 GB/sec
Stdlib Allocate single-threaded scalar -- 1426, , 5.32 Mallocs/sec
Stdlib Write single-threaded scalar -- 1862, , 3.85 GB/sec
Stdlib Copy single-threaded scalar -- 1895, , 1.95 GB/sec

Stream (Score: 1478)
Stream Copy single-threaded scalar -- 1522, , 2.08 GB/sec
Stream Copy single-threaded vector -- 1657, , 2.15 GB/sec
Stream Scale single-threaded scalar -- 1592, , 2.07 GB/sec
Stream Scale single-threaded vector -- 1555, , 2.10 GB/sec
Stream Add single-threaded scalar -- 1183, , 1.79 GB/sec
Stream Add single-threaded vector -- 1640, , 2.28 GB/sec
Stream Triad single-threaded scalar -- 1309, , 1.81 GB/sec
Stream Triad single-threaded vector -- 1371, , 2.57 GB/sec

Geekbench 64 bit scores:

Platform: Mac OS X x86 (64-bit)
Compiler: GCC 4.0.1 (Apple Inc. build 5494)
Operating System: Mac OS X 10.6.4 (Build 10F3061)
Model: MacBookAir3,1
Motherboard: Apple Inc. Mac-942452F5819B1C1B 1.0
Processor: Intel(R) Core(TM)2 Duo CPU U9600 @ 1.60GHz
Processor ID: GenuineIntel Family 6 Model 23 Stepping 10
Logical Processors: 2
Physical Processors: 1
Processor Frequency: 1.60 GHz
L1 Instruction Cache: 32.0 KB
L1 Data Cache: 32.0 KB
L2 Cache: 3.00 MB
L3 Cache: 0.00 B
Bus Frequency: 800 MHz
Memory: 4.00 GB
Memory Type: 1067 MHz DDR3
SIMD: 1
BIOS: Apple Inc. MBA31.88Z.0061.B00.1009101530
Processor Model: Intel(R) Core(TM)2 Duo CPU U9600 @ 1.60GHz
Processor Cores: 2

Integer (Score: 2178)
Blowfish single-threaded scalar -- 1022, , 44.9 MB/sec
Blowfish multi-threaded scalar -- 2051, , 84.1 MB/sec
Text Compress single-threaded scalar -- 1444, , 4.62 MB/sec
Text Compress multi-threaded scalar -- 2484, , 8.15 MB/sec
Text Decompress single-threaded scalar -- 1247, , 5.13 MB/sec
Text Decompress multi-threaded scalar -- 2393, , 9.54 MB/sec
Image Compress single-threaded scalar -- 1379, , 11.4 Mpixels/sec
Image Compress multi-threaded scalar -- 2543, , 21.4 Mpixels/sec
Image Decompress single-threaded scalar -- 1403, , 23.6 Mpixels/sec
Image Decompress multi-threaded scalar -- 2742, , 44.7 Mpixels/sec
Lua single-threaded scalar -- 2413, , 929.2 Knodes/sec
Lua multi-threaded scalar -- 5018, , 1.93 Mnodes/sec

Floating Point (Score: 3377)
Mandelbrot single-threaded scalar -- 1194, , 794.4 Mflops
Mandelbrot multi-threaded scalar -- 2327, , 1.52 Gflops
Dot Product single-threaded scalar -- 2161, , 1.04 Gflops
Dot Product multi-threaded scalar -- 4254, , 1.94 Gflops
Dot Product single-threaded vector -- 1680, , 2.01 Gflops
Dot Product multi-threaded vector -- 3681, , 3.83 Gflops
LU Decomposition single-threaded scalar -- 922, , 820.7 Mflops
LU Decomposition multi-threaded scalar -- 1734, , 1.52 Gflops
Primality Test single-threaded scalar -- 3135, , 468.2 Mflops
Primality Test multi-threaded scalar -- 4504, , 836.0 Mflops
Sharpen Image single-threaded scalar -- 3378, , 7.88 Mpixels/sec
Sharpen Image multi-threaded scalar -- 6346, , 14.6 Mpixels/sec
Blur Image single-threaded scalar -- 4191, , 3.32 Mpixels/sec
Blur Image multi-threaded scalar -- 7781, , 6.12 Mpixels/sec

Memory (Score: 2044)
Read Sequential single-threaded scalar -- 2674, , 3.27 GB/sec
Write Sequential single-threaded scalar -- 2217, , 1.52 GB/sec
Stdlib Allocate single-threaded scalar -- 1580, , 5.90 Mallocs/sec
Stdlib Write single-threaded scalar -- 1838, , 3.80 GB/sec
Stdlib Copy single-threaded scalar -- 1914, , 1.97 GB/sec

Stream (Score: 1593)
Stream Copy single-threaded scalar -- 1436, , 1.96 GB/sec
Stream Copy single-threaded vector -- 1598, , 2.07 GB/sec
Stream Scale single-threaded scalar -- 1510, , 1.96 GB/sec
Stream Scale single-threaded vector -- 1621, , 2.19 GB/sec
Stream Add single-threaded scalar -- 1624, , 2.45 GB/sec
Stream Add single-threaded vector -- 1851, , 2.58 GB/sec
Stream Triad single-threaded scalar -- 1815, , 2.51 GB/sec
Stream Triad single-threaded vector -- 1296, , 2.43 GB/sec