Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

MBAir xBench results :)

Halon X said:
I own Rev A & B MBA's, both with SSD's and for what I use my computer for (email, office, web) see little noticeable difference in performance between the two. If I had to do it again, I would hands down have purchased a Refurb A with the SSD over the B with SSD. If you NEED more graphics horsepower, a Rev B is the clear choice!
Click to expand...
I second that! I have exactly the other 2 (Rev A 1.6 Hdd and Rev B 1.6 HDD) and I feel the same. From XBench you would think the Rev B is 2x faster. It's not even 20% faster, let alone 100%! Only difference is with boot times (70 vs 40 seconds ).
And the YouTube stuttering has been fixed with Flash 10, just download it and it gets much better!
 
AIRniloc, PLEASE add a 150 for the new 2.13 GHz rev B MBA just released. Was reported on another thread by a new user "quisguous." A BIG thanks to you for keeping this thread updated. Perhaps you could change the thread title and sorta reinvent the thread.

XBENCH results for EVERY MBA (now includes the 2.13 GHz SSD rev B MBA v 2,1)!

Thanks!
 
My 2 cents

I have a rev A with the SSD (I put in myself).

It's seems fine. MUCH faster than any netbook (trust me)

However, I run Coolbook and this may be why i don't seem to have any problems.

(Theoretically, a rev. A 1.8GHZ running Coolbook is little different than a rev B 1.8ghz without Coolbook . Coolbook would level the playing field (that's assuming it even works as this is all pretty subjective. But then who cares if the placebo is just a placebo if your happy with the delusion :D )

But i'm sure the rev B is better, i'm just saying.

All n all I fairly like the Air. I've had some of the other Macbooks and the Air is a rather sleek. Also, I prefer the non-glass screen for documents (especially B&W ones). I'm sure the 'glass-gloss' is great for video: it's odd though as the Air's screen is glossy but compared to the glass covered Macbooks it's seems Matte)


whatever
 
One of the greatest differences between the Rev A and the Rev B/C is the change from Intel Graphics to nVidia. Coolbook can't do anything about the graphics.
 
my test

my test score
 

Attachments

  • ProIV128G.png
    ProIV128G.png
    67.6 KB · Views: 84
i plan on updating the first page with the Rev C scores shortly :)

however, has anyone re-run Xbench with SL installed? I just did and noticed a 40 point drop-off :confused::confused::confused:

anyone else experience this?
 
AIRniloc said:
i plan on updating the first page with the Rev C scores shortly :)

however, has anyone re-run Xbench with SL installed? I just did and noticed a 40 point drop-off :confused::confused::confused:

anyone else experience this?
Click to expand...

I read your post and tried, half expecting you did something wrong... as Snow Leopard seems much faster for me. Anyways, you're correct... sorry... and I dropped off from 161 to 125. WTF?

Anyone know what's going on here? Has to be how the program is recording something with SL? I don't have a saved text document from my past tests, but I remembered it being 161. Anyways, I cannot compare to see the discrepancy. The test shows the performance and scores for each part of the test. We should be able to tell what's going on by comparing a before and after SL upgrade. Also, mine was an upgrade. Perhaps that's part of the issue?
 
yeah, i'm running xbench 1.3...

my score in leopard was around 125 on my rev B SSD....now it's hitting in the 80s everytime i perform the test! i don't have my older text files to compare the scores with...

hmmmm

very odd
 
AIRniloc said:
yeah, i'm running xbench 1.3...

my score in leopard was around 125 on my rev B SSD....now it's hitting in the 80s everytime i perform the test! i don't have my older text files to compare the scores with...

hmmmm

very odd
Click to expand...

Well mine dropped from 161 to 125 on a 2.13 GHz/SSD rev C MBA. These are huge drops once SL is installed.

Look to see what segment(s) of the test is lower than previously.

Snow Leopard is obviously changing a component of the test. The drive read speeds are making huge differences in test scores. I would check that first.

Snow Leopard has been really frustrating, but I have to assume this is a problem with Xbench reporting/testing on SL and not that SL is slowing a component of the test down.
 
caonimadebi said:
130.64 on a 2.13Ghz/SSD Macbook Air with a clean install of Snow Leopard.
Click to expand...

That's just the point. Scores are much lower once SL is installed.

I scored a 161 with Leopard and 125, 131, 129 with Snow Leopard. Why such a big drop?

Something has to be going on with the Xbench program while running on SL.

Can someone please post full results of text document with Leopard on a 2.13 GHz rev C SSD? I didn't save old results from Leopard and I need to compare.
 
i'm willing to bet this is an Xbench error...

just at first glance, it seems my CPU and Open GL numbers are WAYYYYYY lower than they used to be...which is even weirder, as these are the two areas which i expected a slight bump in..

oh well...a test is a test
 
im testing on mbp 2.26 and was using geekbench where my score increased with SL
i agree that xbench might be the issue here too
my number is 115 and also, i have 4gb ram and SL installed.
mmmm
 
It's the Open GL test. Uncheck that and you will still get a comparable score. Open GL in SL will give close to half the Leopard score. It's not just an Air thing:

Leopard:
Results 131.32
System Info
Xbench Version 1.3
System Version 10.5.6 (9G55)
Physical RAM 2048 MB
Model MacBookAir2,1
Drive Type APPLE SSD SM128
CPU Test 114.46
GCD Loop 129.53 6.83 Mops/sec
Floating Point Basic 105.81 2.51 Gflop/sec
vecLib FFT 87.95 2.90 Gflop/sec
Floating Point Library 156.15 27.19 Mops/sec
Thread Test 217.93
Computation 351.82 7.13 Mops/sec, 4 threads
Lock Contention 157.86 6.79 Mlocks/sec, 4 threads
Memory Test 169.20
System 182.75
Allocate 190.35 699.01 Kalloc/sec
Fill 160.97 7826.50 MB/sec
Copy 202.02 4172.67 MB/sec
Stream 157.53
Copy 150.52 3108.91 MB/sec
Scale 147.64 3050.26 MB/sec
Add 167.34 3564.75 MB/sec
Triad 166.68 3565.64 MB/sec
Quartz Graphics Test 132.68
Line 130.74 8.70 Klines/sec [50% alpha]
Rectangle 151.99 45.38 Krects/sec [50% alpha]
Circle 115.05 9.38 Kcircles/sec [50% alpha]
Bezier 137.52 3.47 Kbeziers/sec [50% alpha]
Text 133.44 8.35 Kchars/sec
OpenGL Graphics Test 123.26
Spinning Squares 123.26 156.36 frames/sec
User Interface Test 158.57
Elements 158.57 727.75 refresh/sec
Disk Test 82.54
Sequential 64.72
Uncached Write 60.62 37.22 MB/sec [4K blocks]
Uncached Write 56.62 32.04 MB/sec [256K blocks]
Uncached Read 45.38 13.28 MB/sec [4K blocks]
Uncached Read 178.17 89.55 MB/sec [256K blocks]
Random 113.91
Uncached Write 51.03 5.40 MB/sec [4K blocks]
Uncached Write 79.16 25.34 MB/sec [256K blocks]
Uncached Read 1399.78 9.92 MB/sec [4K blocks]
Uncached Read 460.87 85.52 MB/sec [256K blocks]

Snow Leopard w. Open GL:
Results 118.20
System Info
Xbench Version 1.3
System Version 10.6 (10A432)
Physical RAM 2048 MB
Model MacBookAir2,1
Drive Type APPLE SSD SM128
CPU Test 131.60
GCD Loop 218.14 11.50 Mops/sec
Floating Point Basic 106.76 2.54 Gflop/sec
vecLib FFT 85.88 2.83 Gflop/sec
Floating Point Library 208.42 36.29 Mops/sec
Thread Test 246.38
Computation 338.87 6.86 Mops/sec, 4 threads
Lock Contention 193.55 8.33 Mlocks/sec, 4 threads
Memory Test 172.72
System 194.59
Allocate 230.20 845.39 Kalloc/sec
Fill 162.52 7902.08 MB/sec
Copy 203.24 4197.92 MB/sec
Stream 155.28
Copy 149.53 3088.52 MB/sec
Scale 146.47 3025.99 MB/sec
Add 163.86 3490.65 MB/sec
Triad 162.79 3482.56 MB/sec
Quartz Graphics Test 136.64
Line 122.43 8.15 Klines/sec [50% alpha]
Rectangle 148.86 44.44 Krects/sec [50% alpha]
Circle 128.00 10.43 Kcircles/sec [50% alpha]
Bezier 144.81 3.65 Kbeziers/sec [50% alpha]
Text 143.10 8.95 Kchars/sec
OpenGL Graphics Test 57.79
Spinning Squares 57.79 73.31 frames/sec
User Interface Test 185.35
Elements 185.35 850.64 refresh/sec
Disk Test 85.05
Sequential 67.20
Uncached Write 65.13 39.99 MB/sec [4K blocks]
Uncached Write 56.12 31.75 MB/sec [256K blocks]
Uncached Read 47.95 14.03 MB/sec [4K blocks]
Uncached Read 182.12 91.53 MB/sec [256K blocks]
Random 115.79
Uncached Write 51.74 5.48 MB/sec [4K blocks]
Uncached Write 80.87 25.89 MB/sec [256K blocks]
Uncached Read 1439.31 10.20 MB/sec [4K blocks]
Uncached Read 463.67 86.04 MB/sec [256K blocks]

Snow Leopard w.o. Open GL:
Results 141.30
System Info
Xbench Version 1.3
System Version 10.6 (10A432)
Physical RAM 2048 MB
Model MacBookAir2,1
Drive Type APPLE SSD SM128
CPU Test 131.83
GCD Loop 216.93 11.43 Mops/sec
Floating Point Basic 106.38 2.53 Gflop/sec
vecLib FFT 87.87 2.90 Gflop/sec
Floating Point Library 201.98 35.17 Mops/sec
Thread Test 236.73
Computation 267.21 5.41 Mops/sec, 4 threads
Lock Contention 212.49 9.14 Mlocks/sec, 4 threads
Memory Test 172.74
System 193.14
Allocate 224.69 825.15 Kalloc/sec
Fill 162.84 7917.43 MB/sec
Copy 202.38 4180.07 MB/sec
Stream 156.24
Copy 149.59 3089.73 MB/sec
Scale 146.59 3028.58 MB/sec
Add 164.37 3501.33 MB/sec
Triad 166.38 3559.27 MB/sec
Quartz Graphics Test 135.29
Line 120.87 8.05 Klines/sec [50% alpha]
Rectangle 148.19 44.24 Krects/sec [50% alpha]
Circle 127.15 10.36 Kcircles/sec [50% alpha]
Bezier 143.66 3.62 Kbeziers/sec [50% alpha]
Text 140.59 8.79 Kchars/sec
User Interface Test 186.95
Elements 186.95 858.01 refresh/sec
Disk Test 82.49
Sequential 64.14
Uncached Write 42.46 26.07 MB/sec [4K blocks]
Uncached Write 77.35 43.77 MB/sec [256K blocks]
Uncached Read 49.44 14.47 MB/sec [4K blocks]
Uncached Read 176.88 88.90 MB/sec [256K blocks]
Random 115.53
Uncached Write 51.80 5.48 MB/sec [4K blocks]
Uncached Write 80.06 25.63 MB/sec [256K blocks]
Uncached Read 1488.30 10.55 MB/sec [4K blocks]
Uncached Read 464.41 86.17 MB/sec [256K blocks]
 
But why is OpenGL reporting so much slower in Snow Leopard? Why is the critical question. Is it acceptable that SL is off in OpenGL scores? Is it good? Why?
 
NC MacGuy said:
I thought this was well known?

http://www.macnn.com/articles/09/08/29/benchmarks...snow.leopards/

There's more out there but this pretty much summarizes.
Click to expand...

So this is a problem with Snow Leopard that needs to be corrected. What it means is that the xBench scores are correct and until Apple fixes the issue with OpenGL we will see slower performance with Snow Leopard.

I don't believe the xBench test should be run without selecting OpenGL. The whole point is that Snow Leopard isn't functioning as well as Leopard and that is why the xBench scores are lower in SL.

Yet another reason to not upgrade to SL as it appears it isn't ready. I must admit I have been rather disappointed and unimpressed by Snow Leopard and Apple.

It's too bad as I hoped this was a problem with the xBench test somehow and not Snow Leopard.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back