Qualcomm revealed X Elite's benchmark scores

[stinksroundhere]

CPUs should be compared on the basis of CPU cost, but we don't know that. What other metric could we use to compare CPUs? Number of transistors? Die size and node type?

Customers use machines, not CPUs. They don't know what a CPU costs or what Geekbench score it gets. Most customers don't know what a benchmark is and they are not interested in 4K gaming on a laptop.

From a customer perspective they are using a machine comprising of keyboard, form factor, build quality and software. That's why some customers are willing to pay more for better overall design. They want everything to be sturdy, reliable and easy to use.

If the whole experience isn't holistic, well designed, well built and optimised then a faster CPU makes less of a difference. 95% of PC laptops and Windows itself sucks from a design and quality perspective.

Moving to ARM hasn't change the way the Windows PC world designs things.

 

[thenewperson]

Weird to me to see all the comments saying some version of "The M4 smokes this!"

Umm... is there an M4 MacBook I've missed?

Does there need to be one? There is an M4 chip in existence and it smokes the X Elite. Don't know why a MacBook is needed to acknowledge that.

 

[thenewperson]

Apples and oranges. Apple controls all the hardware. Qualcomm based hardware (notebooks, laptops) are produced by different manufacturers (Microsoft, Samsung, Asus, HP etc.) and they all have different marketing strategies (branding, product presentation, time to market etc.). You would have a point if (for instance) Microsoft made all the Qualcomm notebooks, but that isn't the case.

Why does that matter at all? All these companies and all their different marketing strategies are culminating in a worse launch experience than Apple's, that's just a fact. That they've all decided on different strategies doesn't stop them from being compared.

It is like saying "The launch of EV's proves once again that Tesla product launches are much better". IMO there is no logic in a 'many to one' comparison like that.

No, it's perfectly fine to compare a Tesla launch to the launch of its competitors under some unified effort separate from Tesla, if that's how things went down. No one forced all these companies (the many) to form a somewhat unified front to take on a single company (the one).

So @Xiao_Xi is right, Apple launches look much better.

 

[name99]

Windows fanboys who told us not to trust leaked benchmarks and to wait for June 18...where are they now?

Remember for a lot of people this is not about "QC beats Apple". Many people are just not interested in running macOS for whatever reason.
For them, the important thing is "I want something close to as nice as Mac hardware, but running my preferred [ie Windows] OS". And now they have it. We should be happy for them, the world is a better place!

Yes, there are idiots who see this in terms of "how can I use this to rant against Apple", but that is a minority; don't let it sway your understanding of what's really going on here.
What matters is "is it better than my current Windows laptop?" and (soon) "is it better than a Lunar Lake laptop"?

 

[name99]

First and foremost this is a Mac forum. Second, M3 got its ass whooped by Snapdragon X Extreme ultra whatever the name of that chip is in CPU Multi-core performance by almost 50%. For the price of M2. I think this gives ARM Windows fans a lot to look forward to. Not sure why would you want ARM PC but it's an option, I guess.

It's not clear this is an especially useful viewpoint.
In terms of performance, everyone knew the QC part comes it at somewhere between M and M Pro levels, no surprise.

So it's about costs. Is that an especially interesting discussion?
Are these devices exactly equal to MacBook Airs or MacBook Pros that cost more? Well, opinions differ. You can obsess over the MP performance all you like, but for plenty of customers other things matter, whether it's the value of MacOS (and integration with phone, watch, etc); or the single-thread latency (ie snappiness); or AI functionality; or quietness.

Arguing tech has at least some value. Running around screaming "but macs cost too much and I would never pay that" to people who are perfectly willing to pay 10% extra for macOS is just a waste of everyone's time.

 

[name99]

It has been mentioned here that Windows on ARM will remain niche and that the majority of Windows users will remain X86 because they have a large library of software and games that will never be ported to ARM and run with a hit or crash on ARM.

That's the big downside of Windows being able to still run a catalogue of very old titles. Their deep backwards compatibility is great for software preservation but it is not great when they need to move the platform to a completely new architecture.

That's something Apple never had to deal with. They always informed developers and customers in advance that major changes will mean old titles will no longer work.

I think this is unproven.
There are clearly SOME people who have software libraries going back to the Reagan years and who obsess over this sort of compatibility. (Though it's also unclear the extent to which that stuff can't simply be virtualized/emulated...)

But there are also plenty of people who just buy a PC as a tool. They have to use it for work because that's what their IT department supports; or they're just used to Windows from experience; or they just don't care about Apple niceness to pay Apple prices. These people have zero interest in ARM vs x86 battles, all they care about is if it runs the 10 apps that matter to them, and what it costs.

My expectation is that Prism will just get better (and other people may wrap up other emulators, like a SoftDOS and a SoftOS2 and a SoftWindows3.1 just because people like to do these sorts of things. Meanwhile the prices will come down (this initial crop of devices are trying to kinda compete at the Apple level; the next round will ship the really lousy screens and slow flash we expect from $500 laptops).

Of course Intel will try to fight back with Lunar Lake. But overall I'm not optimistic Intel can make it work. They're just too incompetent on too many dimensions to compete with QC (and this is QC! so that's really saying something). The very fact that they can't decide whether they want SMT in Lion Cove or not is just symptomatic - they cannot prioritize correctly, so everything gets done, meaning nothing gets done well.

 

[NT1440]

I think this is unproven.
There are clearly SOME people who have software libraries going back to the Reagan years and who obsess over this sort of compatibility. (Though it's also unclear the extent to which that stuff can't simply be virtualized/emulated...)

But there are also plenty of people who just buy a PC as a tool. They have to use it for work because that's what their IT department supports; or they're just used to Windows from experience; or they just don't care about Apple niceness to pay Apple prices. These people have zero interest in ARM vs x86 battles, all they care about is if it runs the 10 apps that matter to them, and what it costs.

My expectation is that Prism will just get better (and other people may wrap up other emulators, like a SoftDOS and a SoftOS2 and a SoftWindows3.1 just because people like to do these sorts of things. Meanwhile the prices will come down (this initial crop of devices are trying to kinda compete at the Apple level; the next round will ship the really lousy screens and slow flash we expect from $500 laptops).

Of course Intel will try to fight back with Lunar Lake. But overall I'm not optimistic Intel can make it work. They're just too incompetent on too many dimensions to compete with QC (and this is QC! so that's really saying something). The very fact that they can't decide whether they want SMT in Lion Cove or not is just symptomatic - they cannot prioritize correctly, so everything gets done, meaning nothing gets done well.

The world’s logistics and banking systems run on ancient systems…

COBOL is still the basis of most financial systems, to the point that banks throw truckloads of cash at retired engineers to come back and help maintain it.

Windows, at its heart, is an enterprise venture that tries to generate value on the consumer side of things.

I’ve been stating since the M1 came out that MS is going to eventually have to break Windows into a consumer facing modern OS, and an “enterprise” version that supports the AGING infrastructure that enables modern life…

 

[Mac_fan75]

The world’s logistics and banking systems run on ancient systems…

COBOL is still the basis of most financial systems, to the point that banks throw truckloads of cash at retired engineers to come back and help maintain it.

Windows, at its heart, is an enterprise venture that tries to generate value on the consumer side of things.

I’ve been stating since the M1 came out that MS is going to eventually have to break Windows into a consumer facing modern OS, and an “enterprise” version that supports the AGING infrastructure that enables modern life…

Yes they do sometimes run Cobol, but for sure they are NOT using ancient systems, it is superior than any crappy instable cloud solution. I am actually working in that business (banking) and here it’s run on HPE NonStop by far not ancient.
not sure where you get that idea nor spreading this disinformation.

 

[NT1440]

Yes they do sometimes run Cobol, but for sure they are NOT using ancient systems, it is superior than any crappy instable cloud solution. I am actually working in that business (banking) and here it’s run on HPE NonStop by far not ancient.
not sure where you get that idea nor spreading this disinformation.

Obviously anecdotal, but from directly speaking to one such former retiree for a regional bank.

Regarding old OS’s, pretty much any independent shop is running industrial machines off a controller running windows 7 (or XP).

My point was that consumers tend to be the market that needs constant innovation while enterprise (where most of MS’s money is) is shockingly (and worryingly) outdated. MS is in a pickle because x86 simply will not be the path forward for consumers, but windows in some way needs to be able to continue supporting old infrastructure.

Hell, I declined work for a Serve Pro franchise because they wanted me to “replace” their server. It was Server 2003…

 

[Xiao_Xi]

MS is in a pickle because x86 simply will not be the path forward for consumers, but windows in some way needs to be able to continue supporting old infrastructure.

I'm still not so sure. Arm-based SoCs in the PC world still have more to prove. We don't know how Qualcomm's SoC will compete with AMD's new CPUs.

Someone has obtained an engineering sample of one of AMD's newest CPUs and has written about the Zen 5 improvements.
https://blog.hjc.im/zen5-preliminary-review.html (Source, Chinese)
https://wccftech.com/amd-strix-poin...latency-throughput-various-performance-tests/ (Translation)

 

[bousozoku]

Obviously anecdotal, but from directly speaking to one such former retiree for a regional bank.

Regarding old OS’s, pretty much any independent shop is running industrial machines off a controller running windows 7 (or XP).

My point was that consumers tend to be the market that needs constant innovation while enterprise (where most of MS’s money is) is shockingly (and worryingly) outdated. MS is in a pickle because x86 simply will not be the path forward for consumers, but windows in some way needs to be able to continue supporting old infrastructure.

Hell, I declined work for a Serve Pro franchise because they wanted me to “replace” their server. It was Server 2003…

Business languages, such as RPG, COBOL, and PL/I always put the needs of the database integrity first by providing consistent data size.

While IBM has operating systems that point back to the 1980s, they're hardly ancient and they're certainly more reliable than these desktop operating systems.

Apple and Microsoft used a larger machine to run their businesses, the IBM System/38, which had a database under the operating system to provide an unified interface.

Thankfully, Linux servers can be substituted for Microsoft Windows servers.

 

[Isengardtom]

Does there need to be one? There is an M4 chip in existence and it smokes the X Elite. Don't know why a MacBook is needed to acknowledge that.

I think the point is that you can’t buy one right now, so at the moment that M4‘s advantage isn’t there when you need a new computer now.

 

[Xiao_Xi]

Has anyone done a good battery test comparing AMD, Intel and Qualcomm based laptops?

 

[NT1440]

I'm still not so sure. Arm-based SoCs in the PC world still have more to prove. We don't know how Qualcomm's SoC will compete with AMD's new CPUs.

Someone has obtained an engineering sample of one of AMD's newest CPUs and has written about the Zen 5 improvements.
https://blog.hjc.im/zen5-preliminary-review.html (Source, Chinese)
https://wccftech.com/amd-strix-poin...latency-throughput-various-performance-tests/ (Translation)

Please keep in mind I’m talking 5-10 years out. Businesses don’t turn on a dime.

I’m excited to see what AMD pulls off in low wattage (for whatever my next work machine will be), but I don’t see a long term (10+ years) path forward for x86 in purely consumer laptops unless Windows has some type of reboot.

 

[thenewperson]

I think the point is that you can’t buy one right now, so at the moment that M4‘s advantage isn’t there when you need a new computer now.

And those who’d need a computer later, perhaps when an M4 laptop has launched, would appreciate the comparison with the M4, considering it does still exist even if it’s not in a Mac right now.

 

[komuh]

I hope they make their own linux based system (join System76 effort maybe?) and focus on proper power efficiency and remove spyware, it could be true competition to macbooks if they focus on getting better OS as whole package seems to be a better value compared to M3 (Intel and AMD leaks looks even better lets see how cheap high quality laptop can be in next few years).

 

[name99]

The world’s logistics and banking systems run on ancient systems…

COBOL is still the basis of most financial systems, to the point that banks throw truckloads of cash at retired engineers to come back and help maintain it.

Windows, at its heart, is an enterprise venture that tries to generate value on the consumer side of things.

I’ve been stating since the M1 came out that MS is going to eventually have to break Windows into a consumer facing modern OS, and an “enterprise” version that supports the AGING infrastructure that enables modern life…

All true. But none of it relevant to my point about the PCs bought by most CONSUMERS...

 

[falainber]

Has anyone done a good battery test comparing AMD, Intel and Qualcomm based laptops?

Here is one relevant article: Snapdragon X Elite laptops last 15+ hours on our battery test, but Intel systems not that far behind

 

[crazy dave]

Has anyone done a good battery test comparing AMD, Intel and Qualcomm based laptops?

Here is one relevant article: Snapdragon X Elite laptops last 15+ hours on our battery test, but Intel systems not that far behind

The article isn't bad, but as much I was ragging on MaxTech for not recognizing throttling on battery for intense multicore workload he was running on the Snapdragon, the Intel and AMD machines should be worse in this regards. Battery tests are ... hard. Unless you have a ridiculous situation like when the M1 first came out 4 years ago, there needs to be a lot of context with regards to performance and specs.

The NoteBookCheck article @TheSl0th linked to, measuring power draw is amongst the better, where you can see that Snapdragon is very good overall compared to Intel and AMD*, and I've seen but some Chinese YouTubers do full battery tests under lots of different loads. I don't know that they've done them for Snapdragons yet ... (I think they have but searching for that is a bit harder)

That said this isn't 4 years ago. Releasing effectively an M2 in 2024 isn't going to make as big an impact as the M1 did in 2020. Intel and AMD have indeed made gains and will continue to make them later this year. MS wants to make sure they don't lose anyone else to Apple, but Qualcomm obviously wants more than that. Qualcomm needs to convince Windows users to go with them over the competition not just to tempt defectors back or stop further defectors.

*There are some non-CPU tests that can tell a different story though.

 

[varezhka]

I’m also curious how these new Snapdragon X on Windows PC do on standby. Because that was the one big usability improvement with M series MacBooks. I can just close the lid and expect an iPad like standby time. Wintel machines never really got there in real life usage. I know this is both hardware and software matter so Intel is not the only one to blame, but I really hope things will be better with the new chips.

 

[Xiao_Xi]

I’m also curious how these new Snapdragon X on Windows PC do on standby.

Anecdotal evidence shows that it may be very good.

https://twitter.com/i/web/status/1804526212523438279

 

[ChrisA]

MS is in a pickle because x86 simply will not be the path forward for consumers, but windows in some way needs to be able to continue supporting old infrastructure.

So they offer both ARM and X86 versions of Windows. Just like Linux supoert s dozen of so different architectures, all with the same source code.

Actualy Apple is still suporting both X86 and ARM. MS will have to do the same but for a MUCH longer period.

 

[name99]

Anecdotal evidence shows that it may be very good.

https://twitter.com/i/web/status/1804526212523438279

Does anyone know the technical issues involved here?

When claims like this first appeared a week or so ago, the x86 supporters insisted that Windows just sucked, it was all Windows' fault for waking the machine at random times, blah blah. Clearly that appears not to be the case.

The next line of defense is probably going to be something something drivers. Which may even be true? Are drivers on x86 Windows Laptops known (or at least believed) to be incompetently written? Is there a reason they might be better written on ARM Windows? (Maybe simply starting from scratch rather than inheriting a code base from 1985? Maybe there are just vastly fewer rules yo have to keep in mind on ARM while trying to remain legal, while balancing performance and power?)

As a second control cases, how does x86 Linux do in this case? The most vocal defenders of x86 all kinda implied that Linux did not have this problem while not addressing
- does Linux even have the equivalent of App Nap (ie brief wakeups to load in notifications, email, etc) the way Windows and Mac do?
- if it does, what are its implications for "overnight performance with the lid down"?

 

[Xiao_Xi]

When claims like this first appeared a week or so ago, the x86 supporters insisted that Windows just sucked, it was all Windows' fault for waking the machine at random times, blah blah. Clearly that appears not to be the case.

Microsoft seems to have messed up with Modern Standby, and AMD/Intel CPUs did not help to hide the problem.

Windows Modern Standby is bad. Here's a way Microsoft can fix it.

Microsoft's "modern standby" technology in Windows has questionable benefits and does more harm than good for most people.

www.xda-developers.com

 

[crazy dave]

Does anyone know the technical issues involved here?

When claims like this first appeared a week or so ago, the x86 supporters insisted that Windows just sucked, it was all Windows' fault for waking the machine at random times, blah blah. Clearly that appears not to be the case.

The next line of defense is probably going to be something something drivers. Which may even be true? Are drivers on x86 Windows Laptops known (or at least believed) to be incompetently written? Is there a reason they might be better written on ARM Windows? (Maybe simply starting from scratch rather than inheriting a code base from 1985? Maybe there are just vastly fewer rules yo have to keep in mind on ARM while trying to remain legal, while balancing performance and power?)

As a second control cases, how does x86 Linux do in this case? The most vocal defenders of x86 all kinda implied that Linux did not have this problem while not addressing
- does Linux even have the equivalent of App Nap (ie brief wakeups to load in notifications, email, etc) the way Windows and Mac do?
- if it does, what are its implications for "overnight performance with the lid down"?

Microsoft seems to have messed up with Modern Standby, and AMD/Intel CPUs did not help to hide the problem.

Windows Modern Standby is bad. Here's a way Microsoft can fix it.

Microsoft's "modern standby" technology in Windows has questionable benefits and does more harm than good for most people.

www.xda-developers.com

As @Xiao_Xi linked to there were problems with Windows and "Modern Standby" but I'm not sure the state of play now. The issue was a confluence between OEMs not instituting all possible sleep states in their products, Windows making bad assumptions, and power hungry x86 chips. That was a couple of years ago and MS claimed to be working on it. To be fair, Macs could occasionally experience similar issues, just far less often and more easily circumventable (turning off "Wake for Network access" basically means it will never happened on the Mac).

As for Linux ... I think Hector is working hard to make sure sleep states work as intended on Asahi Macs, but a lot of OEMs don't do that and my memory is that sleep states, and power management in general, on Linux laptops has a really bad reputation overall. Here's one recent-ish Reddit thread:

https://www.reddit.com/r/AsahiLinux/comments/1907jp4