Discontinuation of ARM Notebook with Snapdragon X Elite SoC(tuxedocomputers.com) |
Discontinuation of ARM Notebook with Snapdragon X Elite SoC(tuxedocomputers.com) |
I will say that Intel has kind of made the original X Elite chips irrelevant with their Lunar Lake chips. They have similar performance/battery life, and run cool (so you can use the laptop on your lap or in bed without it overheating), but have full Linux support today and you don't have to deal with x86 emulation. If anyone needs a thin & light Linux laptop today, they're probably your best option. Personally, I get 10-14 hours of real usage (not manufacturer "offline video playback with the brightness turned all the way down" numbers) on my Vivobook S14 running Fedora KDE. In the future, it'll be interesting to see how Intel's upcoming Panther Lake chips compare to Snapdragon X2.
No gaming - and I came in knowing full well that a lot of the mainstream programs don't play well with snapdragon.
What has amazed me the most is the battery life and the seemingly no real lag or micro-stuttering that you get in some other laptops.
So, in all, fine for light use. For anything serious, use a desktop.
Many people including myself do serious work on a macbook, which is also ARM. What's different about this qualcomm laptop that makes it inappropriate?
Unfortunately, even Intel is moving in that direction whenever they're trying to be "legacy free", but I wonder if that's also because they're trying to emulate the success of smartphone SoC vendors.
Instead I paid the premium for a nicely specced Macbook Pro, which is honestly everything I wanted, safe for Linux support. At least it's proper Unix, so I don't notice much difference in my terminal.
Depends why the Snapdragon chips were relevant in the first place! I got an ARM laptop for work so that I can locally build things for ARM that we want to be able to deploy to ARM servers.
(I'm keen about ARM and RISC-V systems, but I can never actually justify them given the spotty Linux situation and no actual use case)
That isn't to say that modern standby/s2-idle isn't super useful, because it is, but more for actual use cases where the machine can basically go to sleep with the screen on displaying something the user is interacting with.
One thing that I find suspicious is the large delta in single thread score between ARM and X86 currently. The real world performance does not suggest that big of a difference in actual use. The benchmarks suggest a 25% performance delta but in actual use the delta seems to be less than 10%. Of course Apple Silicon has the efficiency crown very much locked down
Since they have become a marketing target the benchmarks have become much less useful.
I am so grateful to the Asahi Linux guys who made this whole thing work. What a tour de force! One day, we'll get the M4 Mac Mini on Asahi and that will be far superior to this Snapdragon X Elite anyway.
I remember working on a Qualcomm dev board over a decade ago and they had just the worst documentation. The hardware wouldn't even respond correctly to what you told it to do. I don't know if that's standard but without the large amount of desire there is to run Linux on Apple Silicon I didn't really anticipate support approaching what Asahi has on M1/M2.
For all the flack Qualcomm takes, they do significantly more than Apple to get hardware support into the kernel. They are already working to mainline the X2 Elite.
The difference is that Apple only makes a few devices and there is a large community around them. It would be far less work to create a stellar Linux experience on a Lenovo X Elite laptop than on a M2 MacBook. But fewer people are lining up to do it on Lenovo. We expect Lenovo, Linaro, and Qualcomm to do it for us.
Fair enough. But we should not be praising Apple.
They have little experience producing code that is high enough quality it would be accepted into Linux kernel. They have even less experience maintaining it for an extended period of time.
"Linux on Snapdragon X Elite: Linaro and Tuxedo Pave the Way for ARM64 Laptops"
291 points, 217 comments
If you want to change some settings oft[sic] the device, you need to use their terrible Electron application.
My wife is very sensitive to glossy screens and we have big problems to find new laptop for her, as most good ones are glossy now.
Apparently the Windows exclusivity period has ended, so Google will support Android and ChromeOS on Qualcomm X2-based devices in 2026, https://news.ycombinator.com/item?id=45368167
Generally, they are far nicer than Qualcomm when it comes to supporting standard technology.
https://news.ycombinator.com/item?id=45938410
BTW. I don't think Qualcomm SoCs running Windows was just about performance but more of a time-limited exclusivity deal with MS.
My guess is that it's all the same as in Linux phones that they have large blobs of drivers given by the board producer but not being open, but then... Maybe we should invest time in microkernels? Maybe Linux is a dead end because of the monolithic architecture? Because I doubt big companies will change...
Google has already built Chromebooks (which are Linux based) on them, so presumably the necessary drivers exist.
Outside of laptops, NVidia sells its Jetson Devkits and DGX workstations which run Linux and are pretty fast and ARM based.
And System76 also sells a high powered (and $$$) Linux workstation based on an NVidia ARM chipset
So at least for some ARM SOCs, performance issues have largely been solved.
First of all the userspace is completely different, secondly Android throughout the years has been aggressively changing the ways background process work (in the context of Android activities, not bare bones UNIX), thus it isn't the same as GNU/Linux where anything goes.
Google and Samsung managed to make very successful Chromebooks together, but IIRC there was a bunch of back and forth to make the whole thing boot quickly and sip battery power.
For example I've had this dell Elitebook where I've installed Debian wiping out Win. While on windows system prompts Bios update practically every week but been years in Linux on same bios. IIRC updates were win only or jump thru some complex rings of fire. Haven't bothered looking up in a while..
I've also had to disable some protection such as security before I could install Debian though I guess there's a way if I research hard enough.
I was disappointed to see that no more good linux compatible XPS was available anymore because they are now based on the last snapdragon for bullshit windows "ai" reasons.
The hard part isn't the money - it's identifying an addressable market that makes the investment worthwhile and assembling a team that can execute and deliver on it.
The market can't be a few hundred enthusiasts who want to spend $10k on a laptop. It has to be at least tens of thousands who would spend $1-2k. Even that probably won't get you to the break-even when you consider the size (and speciality) of team you need to do all this.
Only RISC-V is worth switching to.
Linux OTOH can only use the information it has from ACPI to accomplish things like CPU power states, etc. So you end up with issues like "the fans stop working after my laptop wakes from sleep" because of a broken ACPI implementation.
There are a couple of laptops with excellent battery life under linux though, and if you can find a lunar lake laptop with iGPU and IPS screen, you can idle around 3-4W and easily get 12+ hours of battery.
It's extremely dependent on the hardware and driver quality. On ARM and contemporary x86 that's even more true, because (among other things) laptops suspend individual devices ("suspend-to-idle" or "S0ix" or "Modern Standby"), and any one device failing to suspend properly has a disproportionate impact.
That said, to a first approximation, this is a case where different people have wildly different experiences, and people who buy high-end well-supported hardware experience a completely different world than people who install Linux on whatever random hardware they have. For instance, Linux on a ThinkPad has excellent battery life, sometimes exceeding Windows.
If you have a dGPU, Linux implementation of the power management or offloading actually consumes more power than Windows due to bad architectural design. Here is a talk from XDC2025 that plans to fix some of the issues: https://indico.freedesktop.org/event/10/contributions/425/
Desktop usage is a third class citizen under Linux (servers first, embedded a distant second). Phones have good battery life since SoC and ODM engineers spend months to tune them and they have first party proprietary drivers. None of the laptop ODMs do such work to support Linux. Even their Windows tooling is arcane.
Unless the users get drivers all the minute PMICs and sensors, you'll never get the battery life you can get from a clean Windows install with all the drivers. MS and especially OEMs shoot themselves in the foot by filling the base OS with so much bloat that Linux actually ends up looking better compared to stock OEM installs.
Think I'm arguing its both things where the OS itself can optimize things for battery life along with instilling awareness and API support for it so developers can consider it too.
This meant that by the time they started pushing devs to pay attention to QoS and such, good Mac apps had already been thoroughly multithreaded for years, making it relatively easy to toss things onto lower priority queues.
Try writing Apple Watch software.
Everything is about battery life.
That's in an extremely vanilla Debian stable install, running in the default "Balanced" power mode, without any power-related tuning or configuration.
That compares reasonably well with my 14" M3 Macbook Pro, which seems to be drawing around 3.5 W with a similar set of apps open.
Sure, the XPS is flattered in this comparison because it has a slightly smaller screen, but even accounting for that it would still be... fine? Easily enough to get through a full day of use, which is all I care about.
There's nothing special about this XPS, and I'd expect the Thinkpad models that have explicit Linux support to be equally fine. The key point is that the vendor has put some amount of care and attention into producing a supportable system.
But that's just one problem, I bet.
I personally never tested it, and I can't find definite benchmarks that confirm and measure the waste.
The comment about ACPI being the problem is slightly off base, since its a huge part of the solution to good power management on modern hardware. There isn't another specification that allows the kind of background fine grained power tuning of random busses/devices/etc by tiny management cores who's entire purpose is monitoring activity and making adjustments required of modern machines. If one goes the DT route as QC has done here, each machine needs a huge pile of custom mailbox interface drivers upstreamed into the kernel customized for every device and hardware update/change. They get away with this in the android space because each device is literally a customized OS and they don't have the upstream turnaround problem because they don't upstream any of it, but that won't scale for general purpose compute as the parent article talks about.
Here is a list of major ARM licensees, categorized by the type of license they typically hold. 1. Architectural Licensees (Most Flexible)
These companies hold an Architectural License, which allows them to design their own CPU cores (and often GPUs/NPUs) that are compatible with the ARM instruction set. This is the highest level of partnership and requires significant engineering resources.
Apple: The most famous example. They design the "A-series" and "M-series" chips (e.g., A17 Pro, M4) for iPhones, iPads, and Macs. Their cores are often industry-leading in single-core performance.
Qualcomm: Historically used ARM's core designs but has increasingly moved to its own custom "Kryo" CPU cores (which are still ARM-compatible) for its Snapdragon processors. Their recent "Oryon" cores (in the Snapdragon X Elite) are a fully custom design for PCs.
NVIDIA: Designs its own "Denver" and "Grace" CPU cores for its superchips focused on AI and data centers. They also hold a license for the full ARM architecture for their future roadmap.
Samsung: Uses a mixed strategy. For its Exynos processors, some generations use semi-custom "M" series cores alongside ARM's stock cores.
Amazon (Annapurna Labs): Designs the "Graviton" series of processors for its AWS cloud services, offering high performance and cost efficiency for cloud workloads.
Google: Has developed its own custom ARM-based CPU cores, expected to power future Pixel devices and Google data centers.
Microsoft: Reported to be designing its own ARM-based server and consumer chips, following the trend of major cloud providers.
These companies license pre-designed CPU cores, GPU designs, and other system IP from ARM. They then integrate these components into their own System-on-a-Chip (SoC) designs. This is the most common licensing model.
MediaTek: A massive player in smartphones (especially mid-range and entry-level), smart TVs, and other consumer devices.
Broadcom: Uses ARM cores in its networking chips, set-top box SoCs, and data center solutions.
Texas Instruments (TI): Uses ARM cores extensively in its popular Sitara line of microprocessors for industrial and embedded applications.
NXP Semiconductors: A leader in automotive, industrial, and IoT microcontrollers and processors, almost exclusively using ARM cores.
STMicroelectronics (STM): A major force in microcontrollers (STM32 family) and automotive, heavily reliant on ARM Cortex-M and Cortex-A cores.
Renesas: A key supplier in the automotive and industrial sectors, using ARM cores in its R-Car and RA microcontroller families.
AMD: Uses ARM cores in some of its adaptive SoCs (Xilinx) and for security processors (e.g., the Platform Security Processor or PSP in Ryzen CPUs).
Intel: While primarily an x86 company, its foundry business (IFS) is an ARM licensee to enable chip manufacturing for others, and it has used ARM cores in some products like the now-discontinued Intel XScale.None of these companies is able to license cores to third parties.
Only ARM can do that. ARM holds a monopoly.
>this is from DeepSeek, ymmv
DeepSeek would have told you this much, given the right prompt. Confirmation bias is unfortunately one hell of a bias.
For the rest of us, what matters is whether we can open digikey/newegg/whatever and buy a few machines and whether they are open enough for us to achieve our goals and their relative costs. So that list of vendors is more appropriate because they _CAN_ sell the resulting products to us. The problem is how much of their mostly off the shelf IP they refuse to document, resulting in extra difficulties getting basic things working.
Right now, it seems like overkill, but not sure what all the health and fitness stuff requires.
The health stuff is really the killer app and they should have pivoted to low power high efficiency use case a long time ago. It doesn't make sense to charge the watch everyday for the little utility it provides.
There's a saying in motorcycling: it's better to be alive than right. There's no upside in being correct if it leaves you worse off.
There are ways to make things better leveraging the Linux way. Make more usable tools for fixing ACPI deficiencies with hotloadable patches, ways of validating or verifying the patches for safety, ways of sharing and downloading them, and building a community around it.
Moaning that manufacturers only pay attention to where their profits come from is not a strategy at all.
It's quite literally a vendor problem created by vendors leading anyone that doesn't run Windows astray in some cases.
If you run Linux, then dare to change your OSI vendor string to "Windows", you've entered into bespoke code land that follows different non-standard implementations for every SKU, where it's coded to work with a unique set of hardware and bespoke drivers/firmware on Windows. You also forgo any Linux forethought and optimizations that went into the "Linux" code paths.
i.e. don't support vendors whose laptops don't work in Linux.
I have an LG Gram 15 from 2021 and it gets 15+ hours under light usage in Linux.
Now I have a Framework. It randomly reboots when I close the lid, the battery life is terrible, etc. I live with it since I like the idea of a repairable laptop.
Everything else around the cpu. apple systems are entirely co-designed (cpu to work with the rest of the components and everything together to work with mac os).
While i'd love to see macbook-level quality on other brands (looking at you, lenovo) tight hardware+software co-design (and co-development) yields much better results.
That still leaves the usual UEFI + ACPI quirks Linux has had to deal with for aeons, but it is much more manageable than (non-firmware) DeviceTree.
The dream of course would be an opensource HAL (which UEFI and ACPI effectively are). I remember that certain Asus laptops had a microstutter due to a non-timed loop doing an insane amount of polling. Someone debugged it with reverse engineering, posted it on GitHub, and it still took Asus more than a year to respond to it and fix it, only after it blew up on social media (including here). With an opensource HAL, the community could have introduced a fix in the HAL overnight.
Apple's hardware+software design combo is nice for things like power efficiency, but so in my experience so far, a Macbook and a similarly priced Windows laptop seems to be about equal in terms of weird OS bugs and actually getting work done.
This is out of the box. With obvious fixes like ripping busted background services out, it gets more than a day. There’s no way normal users are going to fire up console.app and start copy pasting “nuke random apple service” commands from “is this a virus?” forums into their terminal.
Apple needs to fix their QA. I’ve never seen power management this bad under Linux.
It’s roughly on par with noughties windows laptops loaded with corporate crapware.
System76 may be an even better example as they now control their software stack more deeply (COSMIC).
The legend of Windows on ARM is decades old, and people have been seriously trying to make it happen for at least the past two decades. They're all bled dry. Apple is the only one who can turn a profit, courtesy of their sweetheart deal with Masayoshi Son.
If anything, Linux powered devices are a good example on how all of them end up with OEM-name Linux, with minimal contributions to upstream.
If everyone would leave Windows in droves, expect regular people to be getting Dell and HP Linux at local PC store, with the same limitations as going outside their distros with binary blobs, and pre-installed stuff.
Meanwhile Linux is getting a huge popularity boost right now from all the PCs that don't officially support Windows 11 and run Linux fine, and those are distribution-agnostic too because they didn't come with it to begin with.
Usually what is stopping us are the drivers that don't work in other distro kernels, or small utilities that might not have have been provided with source.
Nobody expected Intel to provide employees to write support for 80386 pagetables, or Philips to write and maintain support for the I2C bus. The PC keyboard driver was not sponsored and supported by IBM. Getting the code into Linux was really easy (and it shows in a lot of the older code; Linux kernel quality standards have been rising over time), because everyone was mostly cooperating on a cool open-source project.
But at some point, this became apparently unsustainable, and the expectation is now that AMD will maintain their GPU drivers, and Qualcomm (or some other company with substantial resources) will contribute code and employees to deal with Adreno GPUs. This led to a shift in reviewer attitudes: constant back-and-forth about code or design quality is typical on the mailing lists now.
This means contributing code to the kernel is a massive chore, which any person with interest in actually making things work should prefer to avoid. What's left is language lawyers, evangelists and people who get paid to sit straight and treat it as a 9-5 job.
Which is why most communist like endeavor ends up in failure. Without the necessary pruning that comes with competition, you end up in a situation where it's more profitable to get the power to control the resources and take a fee for each interactions than actually do anything worthwhile to get "rights" to resource allocation.
If you mean HP EliteBooks, it doesn't have to be any more complicated than 1) extract BIOS archive 2) drag & drop BIOS file 3) reboot.
You download the BIOS update .exe, run 7zip on it, and take the BIOS.BIN file and either stick it in the root of your EFI partition and it will install automatically on boot, or just run `fwupdtool install-blob BIOS.BIN` and it will install automatically on reboot.
I believe Elitebooks are Ubuntu Certified, which I would imagine gets their firmware updates pushed to LVFS.
Personally, with the Thunderbolt Dock 4 and an HP ZBook G10, I have gotten timely (<30 days of release) automatic updates from LVFS for both on Ubuntu/Fedora with 0 effort on my part.
Ran fwupdmgr (thanks Google) and it lists a whole bunch of hardware on my laptop as "device with no update found", included in that are UEFI device firmware, UEFI dbx, System firmware etc.
I will check further, thanks again. Didn't know about this.
I'll check on the point you have made, see if I can update bios. Thanks.
I have used Secure Boot with Linux for several years now, too. Microsoft signed the shim loader and most distros can do it out-of-the-box now, much like fwupdmgr (above).
I think the biggest thing is finding, for example, devices that are Ubuntu Certified. You don't have to use Ubuntu necessarily, but the whole ecosystem benefits from hardware manufacturers having a slight degree of accountability having done this.
That's a choice the vendor makes, and Tuxedo Computers is the vendor in this case. Since they control the product they're making, they should be able to provide nice firmware updates for Linux users. They just decided not to even try, I guess?
It's only a "Linux problem" if you're trying to run Linux on hardware that is actively hostile to it. There are plenty of vendors who supply good Linux happy paths in their firmware, using their hardware is the solution to that self-imposed problem.
Native is also much faster than qemu emulation - I have a personal (non-.NET) project where I moved the CI from docker/qemu for x86+arm builds to separate x86+arm runners, and it cut the runtime from 10 minutes in total to 2 minutes per runner.
Outside the embedded space, cross-compilation really is a fool's errand: either your software is not portable (which means it's not future-proof), or you are targeting an architecture that is not commercially viable.
This is what we largely do - my entire team other than me is on x86, but setting up the ARM pipelines (on GitHub Actions runners) would have been a real pain without being able to debug issues locally.
As a point of comparison, I daily two ARM macs (work M4 14 + personal M3 14), and I get far better battery life than that (at least 8 hours of "normal" active use on both). Also, antidotally, the legion of engineers at my office with macs are not seeing battery life issues either.
That said, I have yet to encounter anyone who is in love with macOS Tahoe and it's version of Liquid Glass.
I have macos crash reporting turned off, but crashreport pins the CPU for a few minutes on each ios wallpaper renderer crash. I always have the iOS simulator open, so two hours battery, max.
I killed crashreport and it spun the cpu on some other thing.
In macos 25, there’s no throttle for mds (spotlight), and running builds at a normal developer pace produces about 10x more indexing churn than the Apple silicon can handle.
What really burned me on this kind of stuff was the disappearance of Xeon Phi drivers from the kernel. Intel backed it out after they discontinued the product line, and the kernel people gladly went with it ("who'll maintain this?"). Intel pulled a beautiful piece of process lawyership on it: apparently they could back it out without difficulty, because the product was never released! (Never mind it has been sold, retired and circulated in public.)
If you depend on that hardware, you can get it to be supported again. It just doesn't seem to be all that popular.
Most of the time I am running StumpWM with Emacs on one workspace and Nyxt in another. So just browsing and coding mostly.
OpenBSD gets close, but FreeBSD got a slight edge battery wise. To be fair, that is on an old CPU that still has homogenous cores. More modern CPUs can probably benefit from a more heterogenous scheduler.
One of my 'bad' machines is more like 10-100W and i'm lucky to get two hours.
Smaller efficient CPU + low power sleep + not a lot of background activity + big battery = very long run times.
there needs to be a huge healthy ecosystem/economic incentive.
it's all about the software for end users. I don't care what brand it is or OS and how much it costs. I want to have the most polished software and I want to have it on release day.
Right now, it's Apple.
Microsoft tries to do this but is held back by the need for backward compatibility (enterprise adoption), and Google cannot do this because of Android fragmentation. I don't think anyone is even near to try this with Linux.
Almost everything on regular Fedora works on Ashai Fedora out of the box on Apple Silicon.
You can get a full Ubuntu distribution for RISC-V with tens of thousands of packages working today.
Many Linux users would have little trouble changing architectures. For Linux, the issue is booting and drivers.
What you say is true for proprietary software of course. But there is FEX to run x86 software on ARM and Felix86 to run it on RISC-V. These work like Rosetta. Many Windows games run this way for example.
The majority of Android apps ship as Dalvik bytecode and should not care about the arch. Anything using native code is going to require porting though. That includes many games I imagine.
4% was last year, it was 5% by this summer (a significant YoY increase and about what macOS had in 2010) and the Windows 10 end of support was only last month so the numbers from that aren't even in yet.
> Usually what is stopping us are the drivers that don't work in other distro kernels, or small utilities that might not have have been provided with source.
A lot of these machines are pure Intel or AMD hardware, or 95% and then have a Realtek network controller etc., and all the drivers are in the kernel tree. Sometimes the laptops that didn't come with Linux to begin with need a blob WiFi driver but plenty of them don't and many of the ones that do will have an M.2 slot and you can install a different one. It's not at all difficult to find one with entirely open source drivers and there is no apparent reason for that to get worse if Linux becomes more popular.
I was around when everyone was supposed to switch in droves to Linux back in the Windows XP days, or was it Vista, maybe Windows 7, or Windows 8, eventually 8.1, I guess Windows 10 was the one, or Windows 10 S, nah really Windows RT, actually it was Windows 11,or maybe....
I understand, I used to have M$ on my email signature back in the 1990's, surely to be found in some USENET or mailing list archive, yet we need to face the reality without Windows, Valve would not have a business.
macOS nowadays is closing in on 20%. And you can only buy macOS on premium-priced hardware and by now Linux supports more games than it does. The thing holding either of them back has always been third party software compatibility, which as the web has eroded native apps has been less of a problem, which is why both macOS and Linux have been growing at the expense of Windows.
And these things have tipping points. Can your company ignore Linux when it has 0.5% market share? Sure. Can you ignore it when it has 5% market share? There is a less of a case for that, so more things support it, which allows it to get even more market share, which causes even more things to support it. It's non-linear. The market share of macOS would already be significantly higher than it is if a new Mac laptop didn't start at a thousand bucks and charge $200 extra to add 8GB of RAM. Linux isn't going to have that problem.
Now, is it going to jump from 5% to 50% in three days? Of course not. But it's probably going to be more tomorrow than it was yesterday for the foreseeable future.
> we need to face the reality without Windows, Valve would not have a business.
Valve makes money from selling games and Steam. If Linux had 70% desktop market share and Windows had 5%, what would change about how they make money?