We built our software to take advantage of muticore Ampere. And after several optimizations we were manage to be on pair with Xeon. But the major point was that Ampere Server was twice cheaper than Xeon. AMD Epic was on top beating us by almost 20%, but again the price was the factor.
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.
Microsoft Says Windows Laptops With Snapdragon X Elite Will Be Faster Than M3 MacBook Air
- Thread starter MacRumors
- Start date
- Sort by reaction score
I would rather much like to apple and Microsoft to collaborate on windows bootcamp for apple silicon
Except that:
1) For most tasks, it's single-core performance that matters (because most apps are single-threaded), and it's not clear from the linked article how its SC performance compares.*
2) I believe the Air's performance is the same whether plugged-in or on-battery. By contrast, with its much higher power consumption, chances are on-battery performance of Elite X-equipped laptops is going to be significantly lower than the Air's; and on-battery is a big part of how laptops are used.
So I'm glad there's competition; but I'm not sure how much competition the Elite X represents.
EDIT: I just realized the articles MR is citing are from 2023!
Here's a much more recent article, by Aaron Klotz on Tom's Hardware, that shows both SC and MC scores reported by @techinmul on X (Twitter) (from April 1, 2024). This is just one person's result, so take it with the usual grain of salt, but the reported MC score is 14,254, which is 14,254/11,588 = 23% faster than what Primate reports for the M3 in the base MBP (they don't list scores for the M3 Air). But, by contrast, the M3's SC score is 3,087/2,427 = 27% faster. (https://browser.geekbench.com/mac-benchmarks). [Lazy journalism by Klotz not to include that, and to quote Qualcomm's values relative to the M2 when the M3 is the current chip.]
So you can always get higher MC scores by just throwing more cores in. The question is --- what's the power budget?
https://www.tomshardware.com/pc-components/cpus/snapdragon-x-elite-beats-amd-and-intel-flagship-mobile-cpus-in-geekbench-6-qualcomms-new-laptop-chip-leads-in-single-and-multi-core-tests#:~:text=The%20Snapdragon%20X%20Elite%20scored,Qualcomm's%20new%20chip%20outperforms%20them.
1) For most tasks, it's single-core performance that matters (because most apps are single-threaded), and it's not clear from the linked article how its SC performance compares.*
2) I believe the Air's performance is the same whether plugged-in or on-battery. By contrast, with its much higher power consumption, chances are on-battery performance of Elite X-equipped laptops is going to be significantly lower than the Air's; and on-battery is a big part of how laptops are used.
So I'm glad there's competition; but I'm not sure how much competition the Elite X represents.
EDIT: I just realized the articles MR is citing are from 2023!
Here's a much more recent article, by Aaron Klotz on Tom's Hardware, that shows both SC and MC scores reported by @techinmul on X (Twitter) (from April 1, 2024). This is just one person's result, so take it with the usual grain of salt, but the reported MC score is 14,254, which is 14,254/11,588 = 23% faster than what Primate reports for the M3 in the base MBP (they don't list scores for the M3 Air). But, by contrast, the M3's SC score is 3,087/2,427 = 27% faster. (https://browser.geekbench.com/mac-benchmarks). [Lazy journalism by Klotz not to include that, and to quote Qualcomm's values relative to the M2 when the M3 is the current chip.]
So you can always get higher MC scores by just throwing more cores in. The question is --- what's the power budget?
https://www.tomshardware.com/pc-components/cpus/snapdragon-x-elite-beats-amd-and-intel-flagship-mobile-cpus-in-geekbench-6-qualcomms-new-laptop-chip-leads-in-single-and-multi-core-tests#:~:text=The%20Snapdragon%20X%20Elite%20scored,Qualcomm's%20new%20chip%20outperforms%20them.
Last edited:
You might be right, but Qualcomm should have the skills to produce a much better than average Arm chip. It might not beat Apple on all metrics, but I won't be surprised if it beats the x86 competition for a comparable part and that's a good first entry. Even if it pulls more power than the M3, I doubt it will draw the 120W or whatever the top end Intel laptops draw, so they may get away without throttling when running on battery. It'll lead to shorter battery life, but it sounds like they're trying to emphasize performance.
I think part of why MS is choosing to compare to Apple is to avoid denigrating the x86 chips that they still need to run on for a lot of customers.
I actually happen to have a Pixelbook and an M2 13" Macbook Air sitting right in front of me and, not counting the feet, the Pixelbook is just a few hairs thinner than the M2 Air. I'm quite surprised that Apple did not make the Macbook Air thinner - the Pixelbook has an Intel i5 in it for goodness sake!
For games I bought a refurb Windows laptop that has a dedicated video card. That with my MBA (which I run my life on, exactly to your point above about performance) gives me everything I need and was easier and cheaper than buying a high end Mac. This split works out well. I however don't care for Microsoft's policies on a lot of things so I won't be upgrading to Win 11, I'm fine just leaving it with 10 even without security updates (EOL next year) so I don't have to deal with a Microsoft account.
I found this more optimal than chasing the latest and fastest with only one machine. I was fine with splitting my needs between platforms, I'm assuming others are too.
Do you know that it isn't?
I mean what if they turn down teh performance to the pooint where it just edges the Macbook?
There benefits of EVERYTHING are performance combined with efficiency as it's relative - there is a lot more than just raw performance, or raw efficiency to an argument.
Long way to go before they have the ability to turn down the performance to match Apple. They are still chasing Apple 4-5 years down the track. More likely Apple will just stop throttling and make a big jump with the M4.
How is it that 90+% of the posters here have no idea what they're talking about, and didn't bother to spend even 3 minutes googling for some info? We discussed this chip to death when it was first announced, and at least some of the posters there actually knew what they were talking about.
The X Elite single-core is pretty close to the M3 in the higher TDP configuration - probably about 4% off. OTOH, in the lower TDP config (which is comparable to the Air, definitely doable in a fanless build), it's off my a more meaningful 10% or so. Still not 20%.
As you point out, it's a lot closer to the Pro than the base M3 in MC. But that's not the whole story, as I pointed out in that other thread (see below).
They've also made major strides in multicore efficiency, though they could stand to do more there. There's been a LOT of progress. Saying that Apple has "rested on its laurels" is just massively ignorant.
Battery life will remain a significant win for Apple compared to the higher-TDP config, but the X Elite can run at comparable TDP to the Air. It will lose on perf that way though.
Here's the problem with all these discussions. Everyone talks about this like it's an engineering face-off. It's not. Apple's M3 and M3 Pro are still better, sometimes by a lot, than the X Elite. But that's only on their own terms. The big difference is that the same X Elite can run like a Pro or like a base M3 depending on the power profile. So you can buy one PC that runs like an Air on battery (10-20 hours use), and like a Pro when plugged in. It won't be as good as the Air, but it'll be reasonably close, and it might be better than the Pro for MC work.
Apple's Pro can certainly do that too. But... Apple's Pro chip *costs more money* (and you can't buy it in an Air-sized package). So that's where this will ultimately be decided: in consumers' wallets. If MS can sell this product at a price equal to or lower than the Air, and it can run in both low and higher power modes... it'll be in many ways objectively better than the Air (though not in all ways). It will certainly be more flexible.
So in the end, while the engineering is interesting by itself, the real battle is marketing and pricing.
Glad to see someone not in the 90%. There have been some plausible reports on the X Elite's performance. For example, https://www.xda-developers.com/snapdragon-x-elite-benchmarks/.
The X Elite single-core is pretty close to the M3 in the higher TDP configuration - probably about 4% off. OTOH, in the lower TDP config (which is comparable to the Air, definitely doable in a fanless build), it's off my a more meaningful 10% or so. Still not 20%.
As you point out, it's a lot closer to the Pro than the base M3 in MC. But that's not the whole story, as I pointed out in that other thread (see below).
You can't just take any design, turn up the voltage, and run it at higher clocks. For any given design, they simply will not run past a certain speed no matter how good the cooling is. There are very major changes in both CPU (and GPU) cores in M3, without which the CPUs probably wouldn't run well, or quite possibly at all, at the higher clocks they're using.
They've also made major strides in multicore efficiency, though they could stand to do more there. There's been a LOT of progress. Saying that Apple has "rested on its laurels" is just massively ignorant.
As I wrote, it won't win on SC performance. At least in Windows - Linux does better, and the 4% deficit might turn into a 3-4% win. Not helpful to Microsoft though
Battery life will remain a significant win for Apple compared to the higher-TDP config, but the X Elite can run at comparable TDP to the Air. It will lose on perf that way though.
Here's the problem with all these discussions. Everyone talks about this like it's an engineering face-off. It's not. Apple's M3 and M3 Pro are still better, sometimes by a lot, than the X Elite. But that's only on their own terms. The big difference is that the same X Elite can run like a Pro or like a base M3 depending on the power profile. So you can buy one PC that runs like an Air on battery (10-20 hours use), and like a Pro when plugged in. It won't be as good as the Air, but it'll be reasonably close, and it might be better than the Pro for MC work.
Apple's Pro can certainly do that too. But... Apple's Pro chip *costs more money* (and you can't buy it in an Air-sized package). So that's where this will ultimately be decided: in consumers' wallets. If MS can sell this product at a price equal to or lower than the Air, and it can run in both low and higher power modes... it'll be in many ways objectively better than the Air (though not in all ways). It will certainly be more flexible.
So in the end, while the engineering is interesting by itself, the real battle is marketing and pricing.
No, you're missing the point entirely. They're *engineering* lies. The marketing is a different story - if they price the chip (and thus the laptop) to compete with the Air, they have a good story (though it's not the slam-dumk some are making it out to be).
Apple's early mover advantage on desktop ARM is evaporating, but I suspect they are holding back innovations which will give future M chips massive speed boosts. Apple is designing their own chips so it all goes at Apple's pace, instead of Intel's like last gen.
It is full of bugs and glitches. It has been rock solid until mavericks and then a free fall.
The proof is in the pudding and it's long past time for Qualcomm and Microsoft to either put up or shut up. Don't get me wrong, I'm about as excited as I can be about something I'm definitely not going to buy, but I'm getting real tired of seeing Qualcomm marketing numbers and limited testing under controlled conditions, regurgitated as "faster than Apple Mx" (at the very least it'd be nice if Mac (!) Rumors could stop giving them free PR.)
In the end, if Qualcomm/Windows on ARM succeeds I think it will be a net positive for Apple (and Mac users), while also providing Apple with some much needed competition in the power efficient laptop space. But seriously, until these machines are out in the wild in the hands of unbiased testers/reviewers/users who are not under embargo or other onerous restrictions, it's all just hot air...
In the end, if Qualcomm/Windows on ARM succeeds I think it will be a net positive for Apple (and Mac users), while also providing Apple with some much needed competition in the power efficient laptop space. But seriously, until these machines are out in the wild in the hands of unbiased testers/reviewers/users who are not under embargo or other onerous restrictions, it's all just hot air...
See my edits with more recent data
I still think the x86 comparisons are what's going to matter in the end and Qualcomm is just about matching those x86 processors on single core and trouncing them on multicore. Two interesting things to watch relative to those x86 processors: power consumption and core size. Did they get the multicore advantage by making a bigger chip, or did they prove that Arm is a more efficient use of silicon? And can they achieve those performance numbers with less heat than the x86.
As a ballpark figure, if I've got the numbers right, Qualcomm is using 50% more cores (12 vs 8) to get a 23% multicore performance advantage over M3.
I'll be watching the x86 comparisons more than the Apple ones for now. It's not really competing with Apple, and isn't really giving any insight into what Apple can do with better design (because Apple still has the better design). If it can match or beat x86 at either a lower price point or lower power consumption, then Qualcomm will probably be the first of several companies vying for PC market share with Arm. I continue to hope that their future generations will improve rapidly and help drive the whole Arm ecosystem forward, Apple and all.
Last edited:
All I am hearing here is that this new "Elite" chip and I am putting that in quotes because it's just as ridiculous as Apple's "Ultra Pro Max" naming scheme, what I am hearing is that this chip that could very well be actively cooled will be able to beat the entry level fanless M3 Air that throttles when, and I'll quote some news article, "the hottest core inside the M3 SoC reached up to 114 degrees Celsius".
Color me shocked that an actively cooled chip has better performance than a fanless design. Improving performance in laptops whilst retaining great battery life is something Apple has nailed down over the previous years. What is missing from the M3 Air right now that Snapdragon and Microsoft could improve on? Of course it's good for us consumers if Apple sees some competition, but people buying the M3 Air right now are excited how capable their new Macbooks are and don't see significant flaws that would drive them towards Microsoft/Snapdragon.
Battery life, housing design and quality, pricing and overall usability might be much worse than the Macbook Air.
If Microsoft hardware was on par with the Macbook Air and then offered something on top like a touch screen then in an alternate universe where the Microsoft Surface line didn't turn out awful it might be a solid offering. I won't be touching any Surface laptop/tablet with a ten foot pole, we were a Surface shop managing hundreds of these and if I ever see one again it would be too soon. The Microsoft Surface devices have very high failure rates and I strongly recommend against buying those.
Color me shocked that an actively cooled chip has better performance than a fanless design. Improving performance in laptops whilst retaining great battery life is something Apple has nailed down over the previous years. What is missing from the M3 Air right now that Snapdragon and Microsoft could improve on? Of course it's good for us consumers if Apple sees some competition, but people buying the M3 Air right now are excited how capable their new Macbooks are and don't see significant flaws that would drive them towards Microsoft/Snapdragon.
Battery life, housing design and quality, pricing and overall usability might be much worse than the Macbook Air.
If Microsoft hardware was on par with the Macbook Air and then offered something on top like a touch screen then in an alternate universe where the Microsoft Surface line didn't turn out awful it might be a solid offering. I won't be touching any Surface laptop/tablet with a ten foot pole, we were a Surface shop managing hundreds of these and if I ever see one again it would be too soon. The Microsoft Surface devices have very high failure rates and I strongly recommend against buying those.
So they will run faster. Wow, so much wow.
At the same time, the laptop burns through the table and has fans running so fast that the sound causes cracks in the windows of your house.
Yeah, I'll take the M3.
If they also match Apple's energy efficiency and thus quiet/cool operation with their product, I will of course change my mind.
At the same time, the laptop burns through the table and has fans running so fast that the sound causes cracks in the windows of your house.
Yeah, I'll take the M3.
If they also match Apple's energy efficiency and thus quiet/cool operation with their product, I will of course change my mind.
Whether it is here or not is another matter but efficiency isn't always the case. If I need to get X done is Y time, something that sips power but leaves the job unfinished is useless. If efficiency was the metric by which we choose goods we'd all be using the same products.
Context is important.
Of course they compete with Windows machines! Why do you think Apple almost went out of business in the 90’s? It’s because MS closed the usability gap considerably with Win95 running on superior hardware at 1/2 the price of a comparable Mac.
Admittedly, Apple isn’t as vulnerable today thanks to Steve having built a stronger ecosystem but Apple has been struggling to grow sales for the past couple of years and Qualcomm surpassing Apple’s silicon won’t help.
Good to have competition. The Snapdragon chip will probably run much hotter and the laptops will most likely have much lower battery life than MacBook Air.