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

Kan-O-Z

macrumors 6502
Original poster
Aug 3, 2007
305
2
I thought I would put together some of my findings and to compare the CPU performance of the MBA with each other as well as other Macs :)

Here are the results. The 1.6 MBA Rev B is estimated but it should be pretty close to the real number once someone has tested it.

XBench CPU Scores:

76.93 (MBA 1.6, Rev A, Xbench CPU)
98.95 (MBA 1.6, Rev B, Estimated Xbench CPU)
102.90 (MBA 1.8, Rev A, Xbench CPU)
110.39 (MB Santa Rosa 2.2, 160GB)
115.03 (MBA 1.86, Rev B, Xbench CPU)
120.16 (MBP Santa Rosa 2.4, 160GB)

From this we can conclude a few things. The 1.6 Rev B CPU is pretty much equal to 1.8 Rev A CPU and that the updates to Rev B has increased performance by 8.1% for the exact same clock speed. We can also see that the Rev B 1.86 MBA CPU is faster than MB is nearly on par with the CPU from MBP SR 2.4. Now we can see why these things with the SSDs are fast :) Good stuff!

My next post explains where I got this information from.

Enjoy
 

Kan-O-Z

macrumors 6502
Original poster
Aug 3, 2007
305
2
Here is how I came up with these results.

Here is some test info on the MBA Rev B 1.86 SSD
http://mar247advice.wordpress.com/2008/10/31/special-edition-tech-talk-macbook-air-20-review/

From this we get
http://db.xbench.com/merge.xhtml?doc2=316751

According to this the CPU of the 1.86 measures at:
115.03 (MBA 1.86, Rev B, Xbench CPU)

Now lets extrapolate the Rev B 1.6 CPU from this information:
115.03/1.86*1.60 = 98.95
98.95 (MBA 1.6, Rev B, Estimated Xbench CPU)

Now some info on MBA Rev A 1.6 and 1.8

http://arstechnica.com/reviews/hardware/macbook-air-ssd-review.ars

According to this, the MBA Rev A, MB and MBP CPUs measure at:
76.93 (MBA 1.6, Rev A, Xbench CPU)
102.90 (MBA 1.8, Rev A, Xbench CPU)
110.39 (MB Santa Rosa 2.2 C2D, 160GB)
120.16 (MBP Santa Rosa 2.4 C2D, 160GB)

Kan-O-Z
 

andreab35

macrumors 6502a
May 29, 2008
825
0
USA
Wow... thank you for the information! It is very interesting!
Thank you for contributing so much to the MBA community. :)
 

Kan-O-Z

macrumors 6502
Original poster
Aug 3, 2007
305
2
Not sure I agree. In real world tests the MB 2.2 is still faster:
https://forums.macrumors.com/threads/610234/


Hmm that's interesting. Note the results I posted compare the Air to the SR MB. This one is perhaps the next version up from that so it might be faster?

I heard the 1.86 is downclocked (down throttled) more than the other mac notebooks. This basically means it only runs at 1.86 only when needed otherwise it runs at a lower speed to keep temps down and extend battery life. I wonder if the down throttling could have hurt it here...although it should have ran at 100% in this test. It's also interesting to note that the MB ran hotter than the MBA. The MBA didn't spin up it's fans to max until 4 minutes into the test whereas the MB spun up within a minute. This would support the fact that perhaps the MBA was perhaps not running at 100% for the first 3 minutes or so? Perhaps the throttling on the MBA is a little more conservative meaning it won't jump to full throttle very quickly whereas perhaps the throttling on the MB is more aggressive?

Here is a benchmark that shows the MBA rev B pretty much equal to the current base model white MB:
http://www.macworld.com/article/136760/2008/11/macbook_air_review.html

Either way, I am amazed that the MBA has decent power. When you physically see it and hold it you would never take it seriously in the performance department....but you should. It's a sleeper, esp with SSD :) It's crazy to think that this thing runs faster than 1-2 year old MBPs...which many people still consider to have decent power.

Kan-O-Z
 

knmlee

macrumors newbie
Nov 28, 2008
27
0
Performance compared to old PowerMac G5

I have the new 1.86 MBS SSD and am comparing it to a PowerMac G5 with dual 2ghz processors. Both machines have 2GB ram.

I upgraded to Office 2008 on both machines and ran my financial simulation (Excel) which is a heavy duty spreadsheet and Applescript combo.

The G5 ran the simulation 29% faster than the MBA. Not bad, but I was expecting the MBA to be faster with the Universal app and Intel chip combo.

Oh well.

The really bad news is that my simulation confirms that I'm broke after this market crash (on both machines) :(
 

justit

macrumors 6502a
Dec 1, 2007
640
1
I heard the 1.86 is downclocked (down throttled) more than the other mac notebooks.
Yes, it loves to hover around 800Ghz. I've never seen it sustain 1.86 for more than a second.

...although it should have ran at 100% in this test.

Yes that's what I thought I was paying for.

The MBA didn't spin up it's fans to max until 4 minutes into the test whereas the MB spun up within a minute. This would support the fact that perhaps the MBA was perhaps not running at 100% for the first 3 minutes or so?
Yes, good observation.


Perhaps the throttling on the MBA is a little more conservative meaning it won't jump to full throttle very quickly whereas perhaps the throttling on the MB is more aggressive?

Agreed, if coolbook is updated to run on Rev B you could 'force' it to throttle up. But someone needs to come up with software to gauge throttle based on apps. If I'm browsing, keep it at 800Ghz but if I'm encoding keep it at 1.6-1.86 until its finished. Yes the fans would go crazy but isn't it suppose to manage that kind of clock speed performance?

- my 2 cents
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.