I was excited about this as an upgrade from my 2011 MacBook Air 11", until I read this thread, apparently it sucks, thanks.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
First 2018 MacBook Air Benchmark Shows Moderate Gains Over 2017 MacBook Air and MacBook
- Thread starter MacRumors
- Start date
- Sort by reaction score
I wouldn’t pay too much attention to that if I was you, I’d wait for the official reviews. To me it looks like a great update to the Air.
[doublepost=1541185532][/doublepost]
Yea noticed that, towards the end of the video they do a benchmark test.
There are compromises on those iPads. The construction is cheaper. There is no fully laminated display. If you’re making the same 2013 product in 2018 it will be much cheaper to manufacture regardless of the newer SoC.
This iPad at $329 is what is called a loss leader. The low price is made up for with services. Apps, subscriptions, AppleCare and the like.
Thank you for putting the options into perspective - very helpful.
I believe that, as always, Apple has been very strategic with their pricing and what they have (and haven't) chosen to update (i.e. nTB MP Pros). Very frustrating.
My dilemma is that the price of the new Air ($2149 NZD) is similar only to a 2017 Pro (new $2550 NZD or refurb $2400 NZD) that only has the 7th gen. chip and 2nd gen. keyboard.
The price difference of the Air makes it attractive as it gains the 8th gen. chip and 3rd gen. keyboard, but at what day-to-day performance compromise?
They are not the same. 7W vs. 5W. They are close but not the same. The memory is also running faster at 2133 MHz vs. 1866. To be honest we need to wait to see the numbers. There is almost no info available.
Like I need there’s not a big difference but I expect better performance even if it is not much.
We don't need to wait, Geekbench numbers have been posted. 4340 is the best seen so far which puts it higher than the 2017 ntMBP.
The reason they haven’t updated the base 13” is because they will not include the same quad core i5s in the 13” TB and there is no newer dual core i5 with Iris with lower clock speeds.
Otherwise they would be offering the same TB MBP without the TB and guess which one will sell more.
While I’m glad Intel has given every level 2 extra cores, they really screwed Apple’s product differentiation.
[doublepost=1541187337][/doublepost]
OK. It seems Geekbench does factor in the integrated GPU if any is onboard into their CPU single/multi core scores. I never run benchmark software myself since I care about real world performance and the benchmarks are already out long before I am able to buy any new release.
Edit: I made a mistake and no longer think this is correct. These Geekbench scores are CPU only without considering integrated GPU.
I did used to benchmark desktop GPUs but that was when I was a teenager and in the days I used ATX builds.
Last edited:
Exactly....real world usage is what really matters. I keep referring back to the Geekbench scores because so many people are claiming that it's slower than the 2017 nTB MBP so why would you buy the 2018 MBA? Except it isn't slower. This article is just wrong per the screenshot I provided.
I tried out a 2017 MB with the i7 processor last year. The performance was perfectly good for my usage. That wasn't why I returned it though. I returned it because the design is too small IMO. It was highly annoying to me to pick it up with one hand on either side and end up hitting a few keyboard keys every time. I went back to my trusty 2013 i7 MBA that is finally being replaced with the 2018 variety. I think this new design with smaller footprint and actual places on either side of the keyboard to pick it up will make it the perfect size for me.
Wait you know what? It is impossible the the iGPU is factored into these Geekbench scores.
I know for a fact the Iris 640 (GT3e) is definitely much stronger than any GT2 iGPU. GT3e has double the EUs (48 vs. 24).
So I take what I said back. The Amber Lake core i5 CPU is comparable to the 15w Kaby Lake i5 in the 13” MBP but not including GPU performance. There is just no way this is possible.
They give a break down of everything tested; I don't see why GPU performance (shared or dedicated) wouldn't be factored.
https://browser.geekbench.com/v4/cpu/10565877
I am not so sure. The CPU can perform GPU functions. The issue is that GPUs are simply better at doing these functions. Think about it:
Compare the 2017 12” MacBook and the 2017 13” Escape scores on the first page. If the GPU was being factored in these scores single core performance would be nowhere near the same.
I’m almost 99% sure the Geekbench single/multi scores do not take the iGPU into account.
I was talking about the GPU specs, they are the same.
Also, the 7Y75 has the same 7 Watt TDP option, and in that mode, it also has the same CPU specs (except for the max memory speed).
The GPU is specifically excluded from those tests as they are CPU tests.
https://browser.geekbench.com/macs/431
If they did take the dedicated GPU into account these scores would just not be close. Especially single core performance.
[doublepost=1541189562][/doublepost]
This is exactly what I thought. Which is why I said we have to wait for benchmarks between the UHD617 compared to the HD615. FCPX export times and stuff like that.
My apologies then, I didn't read it that way. Your earlier post didn't make that distinction clear, because of the way it worded the XS Max/A12 "out-geekbenching" most intel based MacBooks, it came across as you were saying it was actually more powerful than those computers solely because of the geekbench scores
[doublepost=1541191412][/doublepost]
The chips are still inherently different, which makes comparing an 8th gen intel chip to the A12X, for example, a non sequitur
What Apple is doing with ARM is incredible, don't get me wrong. Its some wizardry and it's been impressive since the day they revealed the A7. But the problem is people, not saying you, looks at these geekbench scores and thinks "wow why didn't Apple just put in the A12X in their new MacBook Air" or "intel can't even release 10nm chips and here's Apple with the first 7nm chips!" or those kinds of trains of thought. People say those comments because they look at geekbench scores and think they are directly comparable
...and similar "compromises" could doubtless have been made to build the new Retina MacBook Air in at the same sub-$1000 price as the previous model (I'm not even suggesting a cut in price like the iPad, which I'm not sure was a compromise c.f. the original iPad which was comparatively 'boxy' and didn't even have a camera).
So shouldn't the 2018 MacBook Air logically cost less to make than the model it replaced (the basic industrial design hadn't changed for years - this was Apple's opportunity to streamline the design, use the cheaper butterfly keyboard, cut down on physical ports, add more glue)?
The only actual improvement in the spec that you wouldn't necessarily expect from natural progression is the shift to retina display – (a) that's hardly cutting-edge tech these days and (b) the iPad managed to shift to Retina years ago without a price increase.
Thing is, none of these retail prices are determined by the laws of physics - Apple decides what build cost they want to design to and what margin they want to make on each product.
...bingo. As I said in the previous post, the problem with MacOS is likely that Apple don't get such a rake-off on subscriptions. Otherwise, stick MacOS on an iPad (I'm sure Apple have tried it) add a trackpad and you'd have a better MacBook Air than the MacBook Air.
The new retina MBA is not the same product as the old MBA. The chassis is completely new. That is not the case with the 2017 and 2018 iPad which is very much the iPad Air 1 from 2013. It may look similar to the 13" MBP but it is wedge shaped. The display is exclusive to it as well. It is a 13" Retina display but without P3 color. I would guess it is similar to the early-2015 13" MBP display. It also has a keyboard currently exclusive to the MBA. I would expect the 13" Escape MBP to have it next year but until then they are only making it for this machine.
I would also imagine that this MBA has a different logic board design to the 13" MBP and that again costs money.
I also don't believe the butterfly keyboard is cheaper than the older design.
[doublepost=1541194285][/doublepost]
There is definitely a team that has been trying this for a while now. I'm sure they work on all sorts of stuff that may or may not ever see the light of day.
I don't think this is why Macs cost more money though. Remember it is much cheaper for them to license ARM to make their custom A designs than it is for them to buy a bunch of custom Intel logic boards and CPUs. PC OEMs use a much more standardized logic board design.
Exactly. That's also how you know that Apple is going to ditch Intel chips soon, so they can start using their own A-series chips which are far superior.
I wouldn't be surprised if this happens next year as they are supposedly rolling out the "universal apps" feature which allows devs to build an app once and it will run on all of Apple's devices.
Using the same architecture for their chips in every product would make that possible...
Honestly it'll do fine for office work, browsing, music/ iTunes movies and the like, and that's really it's intended purpose. If your needs are more demanding (moderate video editing?) the nTB Pro offers a good chunk more power for roughly the same money (particularly if you take advantage of the near continuous small discounts on the pro). Arguably the pro is the better value for most people, unless you really want the wedge design, touch ID and (assuming 12h is accurate) ~3-4h more light usage battery life (this is probably largely negated if you're using the machine even moderately anyway as when you start to push it, I'm sure either machine will start depleting a lot more rapidly).