We now have some actual game results now and it seems even worse than the 50% drop seen in the original review.
Instead of 50% of the performance of the 320M, we now have:
26% at a lower resolution in Wow
34% for Lost planet
Those numbers seem to suggest the ULV SAndy Bridge has even worse graphics performance than the previous generation Nvidia 9400M
Gaming performance. Not graphics performance. Don't confuse the two.
Engadget's review said the Intel IGP made short work of 1080p HD clips, so regarding pushing pixels (that aren't games) the HD 3000 seems on par at doing that as the 320m.
Also, I'd venture to day the HD 3000 graphics drivers are more advanced in OSX than they are in Windows.
The same mysterious drop in Gaming performance was seen in Windows vs. OSX in the
Anandtech review of the 13" 2011 MBP:
"Under OS X, the new HD Graphics 3000 GPU is actually about the same performance or even faster than the 2010 13-inch's GeForce 320M. Remember that Apple does a lot of its own driver writing under OS X and the SNB GPU received some TLC from Apple in the form of very well optimized drivers."
And yes, I know the MBP uses a fully clocked IGP and the MBA probably won't.
But if even a fully clocked IGP sucks in Windows and works almost on par with the 320m in OSX, then I'd like to at least see the LV HD3000 benchmarks in OSX before making a final judgement.
it is only 29min. Not sure where you get almost 1 hour from. And it was measured in Windows, so I think this is the most comparable number. Mac OS is known to be better at using less power than Windows. From this, I'd say there would be a marginal increase in battery life by switching to Sandy Bridge - nothing major.
Hah. My bad. I was adding like adding, and not like time adding.
I'd take even a marginal increase in battery life, though, who wouldn't?
And I'd also venture to say that Apple's doing better at battery life than most other manufacturers. The 13" 2011 MBP added 10W to it's TDP and (like you say below) Sandy Bridge seems like it's sneaky with it's turbo boosting - and still the 2011 MBP gets better battery life than it's C2D+320m sibling from last year.
TDP is not the whole story .. for example the 2011 i7 2.3Ghz Sandy Bridge Quad Core is supposed to have a TDP of 45W, which is 10W more than the i7
2.66Ghz 2010 model. However, Anandtech measured the 2011 machine using almost 40W more running a CPU intensive task. Something is very weird about the Sandy Bridge TDP numbers.
Hmm, interesting, I hadn't seen
that comparison yet.
The GPU must come into play in both those test, however... so 45W + 25W = 70W out of the 93W used are accounted for in TDP.
And the 13" MBP pulls 48W instead of it's 35W TDP. It's interesting.
I wish we had seen comparable numbers for the current MBA. Does it pull more than advertised under load? How much? If not, why not? Is turbo boost to blame?
My point was based purely on TDP and not high-end scenarios, the battery life should be longer. Wireless web surfing is how Apple measures it now - and I couldn't see the SL9400/9600+320m combo posting better battery numbers in a wireless web test than the i5/HD3000 combo? That leads me to say unless one was doing high-end Rendering with their MBA - the general web-surfing, itunes playing, facebook-checking Mac user will not see anymore than the ~20W TDP come into play, giving that user longer battery life, no?