Linux on the Arm-Based ThinkPad X13S: Getting There
theregister.com
theregister.com
There's no need to make up some wild conspiracy theories with Windows here, spend a single Google search before going into this direction next time.
I have seen significant changes in the policies established only just weeks ago that should enable Qualcomm to better participate in developer communities. There's a clear message from the CEO and several VPs that these changes are necessary.
Forgive me the snark, but I'll believe it when I see it. The last years didn't inspire me with confidence.
The problem is, at the core, that this sort of behavior used to be the norm in the industry. Working with embedded systems is pure and utter hell. The open documentation and standardized interfaces one got with Intel/AMD is the complete exception.
Indeed, it will be hard to change the minds of all the folks at Qualcomm who have operated here for years under the status quo. I see new leverage for those who have lobbied for change in the past.
I'm cautiously optimistic.
Qualcomm let Apple steal it.
Qualcomm's competition isn't Apple. It's Intel and AMD. Intel and AMD offer chips that run Windows much better than whatever Qualcomm can make. Apple managed to do that because they control the whole stack - they made macOS run well on ARM, something Microsoft has little incentive to do because chicken and egg.
For Windows, binary compatibility with 3rd party apps is extremely important, at least for the apps that aren't browsers running JavaScript, so, what Qualcomm could aim to do is to optimize the heck of their ARM offerings for x86 emulation/JITing, and efficient JavaScript execution, more or less what Apple did.
I found this remark by HN user dwaite very informative: https://news.ycombinator.com/item?id=37327938
TL;DR emulated x86-32 performance on ARM is poor, and is going to stay poor because of the memory model. Whether this strategy is going to be successful largely depends on how important/prevalent legacy 32bit apps are.
So not very exciting, unless Qualcomm and Lenovo are planning to release refreshed hardware soon.
Chicken and egg problem. Qualcomm won't invest massive resources to make a world beating laptop chip if there's no userbase for it.
I think there should be clear incentive there for a CPU manufacturer to establish itself as the reference ARM PC platform provider like Intel is on the x86 side.
Interest for Asahi only proves there's interest for Linux on Apple devices, because they're already so popular and numerous, bought by millions of users first and foremost to run MacOS, not Linux.
There's zero interest for non-Apple ARM laptops because, unlike Macbooks, they have no market share. Nobody's gonna spend hundreds of hours reverse engineering GPU drivers for an ARM laptop that only 20 people on the planet bought, and in half of those cases it's sitting gathering dust.
Like I said, chicken and egg problem. There's not gonna be a massive upsell of ARM laptops if the equivalent SW ecosystem of Apple isn't there to move the HW off the shelves.
Not many are laptops thought
The context of this thread was about ARM laptop devices, not generic ARM widgets. Please use the strongest interpretation of an argument when responding, not the weaker one for the sake of nitpicks, as per HN rules:
>Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.
The software stack is 90% identical and we are seeing smartphones & tables getting closer to laptops both in terms of functionality and form factor
See for example Samsung Dex, "pro" and foldable tablets, chromebook Linux and android subsystem, ...
If it survived an impact with the wall
Apple and MacOS ecosystem guarantee amazing sales and mainstream popularization of these devices, giving them a numbers advantage, meaning it's by far the best choice platform to target with the biggest RoI for your reverse engineering effort versus some more obscure ARM laptops which could be better on paper but nobody bought.
There is a tablet market waiting for such chip. Apple is making a killing with their iPads. I am sure other tablet makers would like a slice of that sweet apple pie.
You're missing the point.
The fact that Android and Windows ARM tablets aren't "making a killing" and not selling even remotely as good as iPads dealing with an abysmal market share, is not because they don't have M1 chips, but it's their shit SW ecosystem that's been a clusterfuck from day one, angering consumers who expected more than just stretched phone apps and 2 years of SW updates, and the indifference and lack of commitment Google has shown to Android tablet ecosystem over the years.
iPads were outselling every other tablets even before they got the M1 chips. In this tablet space it's the available SW ecosystem that sells the HW, not the other way aground. Android tablets could get fictional M3 chips from the future and they still wouldn't outsell iPads because both developers and consumers already got burned by the Android/non-Apple tablet SW experience and decided to stick with Apple.
The whole stack is objectively worse - slower, with inferior software. They are cheaper though, but it doesn't help much when everyone competes in price for what is, essentially, variations of the same product.
There is a good reason why there aren't many Windows ARM tablets and laptops - they perform worse than the low-end Intel stuff. The tooling to make software for them is not great, and the channels through which one distributes the software is also a limitation. In laptops, Windows on ARM needs to compete against Windows on x86 and, unless Qualcomm and Microsoft pull a miracle, those machines will remain uncompetitive.
And yes: apple made a killing from day one. They always had an excellent software stack and a thriving high-quality, ruthlessly curated app ecosystem, something not even Microsoft can't offer with their own products (Hi, Teams team. I'd like to discuss a couple things).
Unlike with PCs, nobody cares how fast their tablet CPUs are when making a purchasing decision as most flagship mobile chips in the last 3-5 years from both Apple and Qualcomm, are more than fast enough to provide a smooth experience for what most people do on a tablet: browse the web, watch Netflix, shop on Amazon, read books, emails and social media.
But consumers and developers gave up on the Android tablets a long time ago because the ecosystem was an abandoned mess, app quality and variety is meh at best, and the SW update support period is far too short, regardless of how powerful the CPUS in those tablets might be, it's the SW that dictated the purchasing decision and what earned the iPad the monopoly market share.
This has a name, "planned obsolescence".
This is why Linux laptop experience improved significantly a few years back when they started disabling unused HDMI ports etc
Erm... A non-Apple ARM laptop you mean.
But yes - these things feel like Celeron laptops for Xeon prices.
The Qualcomm powered Chromebook is the outlier with noticeably worse performance.
It all depends on how and where your power adapters are scattered in your office. I definitely like not worrying about that with the Apples while, with the Dell, I'm constantly mindful about where the adaptor is (and carry one in my backpack).
The confirmed upgrades to 8 performance cores (from 4), ARMv9, PCIe 4, DDR5 RAM and possibly a higher TDP will bring significant performance gains with Gen 4.
But in principle I agree to you, I wouldn't use it yet.
Edit: Also, why leave AMD out :) https://www.cpubenchmark.net/compare/4774vs4104vs4794vs4102/...
I recently went on a work trip with my Pinebook Pro as my only computer, everything worked fine.
I wonder if they plan on making a motherboard replacement with a connector compatible with a Pi CM4 (so I could plug their fine Pine64 SOQuartz, or the Banana Pi CM if I need more CPU, or an 8GB CM4 for the extra memory).
I’d appreciate, however, a bigger screen.
Not a glowing review.
This seems to describe every arm device that isn't a server, or a Raspberry Pi. The mac is in the same situation. It works with a long list of exceptions, given an Asahi image, and pretty much not at all otherwise.
These community projects are great, but by the time Linux gets full support the HW is usually obsolete, replaced with a newer version, or no longer for sale, and there remains a laundry list of won't fix bugs because the hackers have moved on to the next shiny.
I don't understand why this has been going on for > 10 years and arm hasn't bothered to step in and mandate some HW or firmware guidelines like they did in the server space. All these half baked and abandoned devices can't be good for their brand, and its pretty obvious Qualcomm is happy with the current model of milking a string of failures.
No Thanks. (It's the company which knows nothing but planned obsolescence)
For me and I think a lot of others that's one of their main features over the competition.