At a given price, why do Snapdragon laptops offer lesser performance than x86 counterparts?
How is the compatibility with programs obtained from the high seas? (iykyk)
What advantages do Arm chips have over x86 other than battery life?
P. S. I'm considering getting a laptop in the near future. Please answer soon.
White paper here:
In our testing of retail systems, the Snapdragon X2 Elite Extreme produced the fastest CPU results we have measured in a premium thin and light Windows laptop, nearly a 2x advantage in measured AI performance, and leadership that holds in the default Balanced mode and on battery, at a retail price $600 below the tested Intel flagship. Those results describe the three systems we tested.
Each is one implementation of its silicon, and other systems built on the same processors will land differently as OEM design choices vary. The closely matched class and design specifications across these three are what give the comparison its footing, and they are why we consider it fair, representative, and worth drawing conclusions from. the value leader is a rare combination in this market, and it is the combination this platform now holds. The competitive landscape has definitely improved.
Panther Lake brought Intel the best integrated graphics in the segment, and buyers who prioritize integrated gaming may have a real reason to choose it. On every other dimension that defines premium thin and light usage, CPU performance, AI performance, productivity, and behavior on battery, the X2 Elite Extreme leads, and it leads by margins that are unusual in a market accustomed to single-digit generational gaps.
Generation one made Windows on Snapdragon viable and competitive. In our testing, generation two makes it the performance leader in the premium segment. The value position may matter as much as the performance position. The X2 Elite Extreme gives OEM partners a credible flagship at $1,600 to $1,700 price points while comparable Intel X series systems look like they will land $300 to $600 higher. A performance leader that is also The battery life results reinforce the mobility story, with the longest video playback endurance in our testing, the most runtime per watt-hour in every rundown, and a single raw-runtime loss to an Intel
Has anyone purchased open-box A16 from Bestbuy? Anyone know if Asus warranty started with original purchase or if it starts with my purchase? Is purchasing an open box A16 a bad idea?
With the recent price increase to 1995 for this laptop, is it likely to come down anytime soon, or with ram prices being unstable, likely to go up more? (any insider knowledge)
I wrote this as a reply to a previous subreddit post, and I wrote so much that I thought I'd make it a main post. I updated our office to all new computers. Bought many laptops.
My Laptop Experiences: Snapdragon Wins, Others Mostly Fail
Asus X2 Elite A14 (X2 Elite)
The most flawless laptop I’ve used. Three issues:
- External video output hard-capped at 144 Hz (fix this, please) and internal screen is 60hz.
- Limited ports.
- CPU performance drops badly under emulation layers (feels like a 10-year-old chip when the emulator interferes—mainly noticeable in gaming).
Not really a hardware fault on the last point. I’ll buy the A16 when the price drops. ARM + NVIDIA should be excellent (no 144 Hz limit, better GPU headroom), though CPU emulation might still suffer. Qualcomm needs to prioritize GPU performance. Give us a motherboard combo with unlocked frequencies and voltages too. Support DIY.
--------
We bought the ideal port layout for the next laptop: 2× USB-C, 2× USB-A, HDMI, and a headphone jack. Thanks Lenovo:
Lenovo Slim 5x !(custom OLED + battery)! (X2 Plus)
Even better Snapdragon machine—best laptop you can get for ~$900. Perfect for my mother; only real complaint is limited GPU performance. Wish the A14 offered the same customization. Sad it only ships with the 6-core Plus instead of the 10-core option. Fan pitch is a bit high and can squeal. Drop the Extreme chip in and quiet the fans and it would be the best laptop of all time.
The rest were plagued with problems
- MacBook Air M5: MacOS is garbage (enough said, would take too many words). I love classic Macs from the 1990s and iPhones, but after the awful AI, locked-down features (tethering, external monitors, USB-C), and price hikes, my iPhone 17 Air might be my last Apple product. Unsubscribed from all Apple YouTubers. Apple's greatest remaining advantage is a smooth update process thanks to limited hardware SKUs.
- Lenovo Legion 5i (RTX 5050): Great OLED 165 Hz panel, but fans spin >2000 rpm constantly. Windows update bricked the NVIDIA drivers on a brand-new machine—had to run DDU and reinstall twice. Hate Legion/Vantage software almost as much as Asus Armoury Crate.
- Lenovo Slim 7i AURA Intel Ultra 7 Series 3 (355): Almost no ports (just USB-C), and those ports were buggy. Ran poorly until I manually pulled drivers via Intel Support Assistant that Lenovo/Windows never installed. DisplayPort and charging were also flaky. Returned. Might have kept an X7 version.
- Asus Vivobook: HDMI port limited to HDMI 1.4—thirteen years after HDMI 2.0 launched. Same lame restriction as the Snapdragon 144 Hz video-out cap. Single channel ram out of the box slowed it way down. Returned.
Really great powerful processor. I was able to compile code really fast and run games pretty well.
The big issue is I had in the space of three days of heavy use 5 BSODs. I did recommended troubleshooting (I work in IT operations). Drivers, SFC/DISM, and even tried disabling a few settings like memory integrity. Nothing worked so I'm returning it.
I did generally like the device I think I got a bad unit. I also realized it didn't have enough ram for me. I need at least 24 GBs. I though 16 would be fine for my use cases, but emulation of x86 apps uses a lot.
I am thinking about getting an A16 instead, but also looking at intel X7 laptops.
Hey everyone,
I’ve been testing out my ASUS Zenbook A16 (X2 Elite Extreme), and something has been driving me crazy. Even when I’m sitting in "Best Power Efficiency" and Battery Saver mode with very low screen brightness and literally nothing open in the background, the estimated battery time remaining looks completely underwhelming (sitting around 5 hours at 62%, as seen in the screenshot).
For a chip that’s supposed to be an efficiency beast, seeing numbers like this feels super disappointing at first glance.
Is Windows 11’s built-in battery estimation tool just notoriously outdated, terrible, or outright bugged on ARM? Or are other Snapdragon users seeing these wildly inaccurate estimates too, while the actual battery life in real-world usage turns out much better?
Would love to hear your experiences!
Hey yall, I want to buy the ASUS Zenbook A14. Only thing I'm worried about is the Snapdragon chip and if it can run the things I need. I've done my own research but I've never really found a clear answer. I will be using it for some schoolwork, mainly just Microsoft Word. I'm also going to start programming, starting with web development and then Python or some low-level coding. I also take photos so I need it to be able to run Adobe Lightroom. I want it to be able to run Minecraft. That's pretty much it, if there is anything else I should know please let me know and thanks in advance!
This project started with a stupidly simple question:
How far can I push a smartphone if I stop treating it like a smartphone?
This is my Nubia Z70 Ultra with 24 GB RAM and the standard Snapdragon 8 Elite.
The phone is now literally sandwiched between TWO desktop CPU coolers, mounted inside a custom acrylic case. 😂
Yes. Two full-size PC CPU coolers. On a phone.
And surprisingly, it works extremely well.
The interesting part: I'm NOT using one of the newer boosted chips
My Z70 Ultra uses the standard Snapdragon 8 Elite, with its prime CPU cores running at up to 4.32 GHz.
That's important because some of the newer devices I'm comparing against are using newer and/or higher-clocked Snapdragon variants.
For example, the REDMAGIC 11 Pro uses the newer Snapdragon 8 Elite Gen 5, reaching around 4.6 GHz on its prime cores.
So I'm not trying to beat these phones with the same CPU and better cooling.
I'm taking an older, lower-clocked 4.32 GHz Snapdragon 8 Elite and seeing how far it can go simply by giving it completely unreasonable cooling. 😂
In my 3DMark ranking screenshot, my Z70 Ultra reaches 10,533, putting it right among much newer devices and even slightly above the Galaxy S25+ result shown there.
And sustained performance is where the cooling really becomes interesting.
3DMark Wild Life Extreme Stress Test
Best loop: 6,980
Lowest loop: 6,910
Stability: 99% over 20 loops
That's only a 70-point difference between the best and worst loop.
But benchmarks aren't actually the main goal of this project.
I'm trying to replace my PC with the phone
I now have a native Termux + XFCE4 Linux desktop running directly on Android.
External monitor, keyboard, mouse, Linux applications and hardware-accelerated graphics.
I can boot into my desktop environment and use the phone almost like a small ARM computer.
Dofus is running directly from my XFCE desktop, rather than simply mirroring an Android game onto the monitor.
And then there's Windows gaming.
The Witcher 3 — 1080p, ULTRA graphics preset
I'm currently running The Witcher 3 at 1920×1080 using the ULTRA graphics quality preset.
Not Low.
Not Medium.
Not a 720p performance test.
1080p + Ultra graphics preset.
Depending on the scene and GPU load, I'm currently getting approximately 20–30 FPS, with the GPU basically sitting at 99% utilization.
I've even seen it go slightly above 30 FPS in some scenes, as shown in the screenshot.
Obviously, this isn't desktop-GPU performance, but that's not really the point.
This is The Witcher 3 running at 1080p Ultra at around 20–30 FPS on a Snapdragon 8 Elite phone. 😂
And underneath all of this, it's still Android.
So my end goal is to have one device that can do everything:
**Android phone / gaming device
native Linux XFCE desktop
Windows applications
Windows PC games**
The software side is still a work in progress and I'm continuing to optimize everything.
But apparently the solution to smartphone thermal throttling was very sophisticated:
one CPU cooler wasn't ridiculous enough, so I installed two. 😂
I read latest Google books are intel based, what is SD current strategy?
It is clear that it's struggling with sales due to compatibility issues in the windows ecosystem, so why they are not teaming up with aluminium OS? it would provide a perfect match with android apps for desktop use.
It would boost its sales a lot but still that's not happening, even an x1-26-100 would be great compared to current core ultra configuration.
What am I missing?
The X2 Elite was announced last September and we still don't have a mini PC or laptops with 64GB of RAM available? It's been nearly 11 months! Yes, there is a RAM cost issue, and there has been an ASUS mini PC announcement no pricing or other info yet), but seriously, the execution of getting products into peoples hands is so poor - It tooks 6'ish months from announcement of X2 Elite to the availability of the first laptop. It would be much better to delay the product announcement rather than it being 6 months before a device is available as people are getting pissed. Rumour has it of a tweaked X2 chip in September, and we still can't find X2 product with last years announced chip, so disappointing. I was really hoping for a new laptop or mini PC with 64GB+ RAM after they promised devices with 128GB RAM.
Does anyone have insight into additional product availability with higher end specs? Seriously, I am keen, was all set to buy something serious after last years X2 Elite launch event but so little product available with so little variance in spec. <rant over>
A very normal day for me: 20 tabs open, Figma running, references everywhere, a half-finished prototype, and three things I need to export. 😅
That’s pretty much what my design workflow looks like most days.
I’ve been doing all of this on the ASUS Zenbook A14 powered by Snapdragon X2 Elite without breaking my creative workflow, so here’s a quick look at how I go from research and references to the final design.
#MadeOnSnapdragon
Odin 3 UFS 3.1 vs UFS 4.0 – Is this downgrade actually something to worry about?
I recently ordered an Odin 3 Max, and I’ve been following the whole UFS situation.
From what I understand, earlier Odin 3 units were using UFS 4.0, while newer batches are now shipping with UFS 3.1 due to supply issues. AYN also confirmed to me that they couldn’t guarantee/send me a UFS 4.0 unit.
I know that, on paper, UFS 3.1 is significantly slower than UFS 4.0, but I’m trying to understand how much this actually matters for gaming.
Most PC games still work perfectly fine on SATA SSDs, and even demanding games generally don’t seem to require storage speeds anywhere close to the maximum sequential speeds of UFS 3.1. So for current games that the Snapdragon 8 Elite is realistically capable of running, I’m struggling to find a scenario where UFS 3.1 would actually become the bottleneck.
Obviously UFS 4.0 could improve installation times, file transfers and potentially loading times, but would it realistically affect FPS, stuttering or asset streaming in games running through Winlator/GameNative/GameHub?
My bigger concern is future-proofing. I bought the Max model, and it’s a little frustrating knowing that the hardware was originally designed/shipped with UFS 4.0 and was later downgraded to 3.1.
Has anyone actually compared an UFS 3.1 Odin 3 directly with an UFS 4.0 unit?
I’d especially be interested in real-world gaming tests rather than storage benchmarks. If both consoles are running the same demanding PC game with the same settings, is there any measurable difference beyond loading/installing/transferring files?
Basically: is UFS 3.1 genuinely limiting the Odin 3 in any realistic gaming scenario, or are we worrying about a difference that the Snapdragon 8 Elite will never really be able to take advantage of anyway?
Parents with tech-savvy teens, this one's for you 👀
The Snapdragon Insiders team is looking for content creators to review the HP OmniBook 3 powered by Snapdragon X as part of our Access Program.
We're particularly interested in applicants who have teenagers at home who can help put the laptop through its paces with things like:
No, your kids won't be featured in any content. We're simply interested in genuine feedback about how the PC performs in real-life scenarios.
The selected Snapdragon Insider will get to test the device and share their experience with the broader Snapdragon Insiders community.
📅 Applications close August 20.
Apply here: Apply to the HP OmniBook 3 Opportunity
Happy to answer questions below!
I will keep trying....
Ciao ragazzi, ieri ho acquistato un Samsung Book 4 Edge con Snapdragon X Elite. Ieri l ho usato tutto il giorno accendendolo e spegnendolo varie volte. stamattina l ho usato per qualche ora, poi messo in carica fino all 80% poi staccato e spento. dopo circa 6 ore l ho riacceso e ho notato questa specie di artefatti sullo schermo. sono sicuro al 100% che fra ieri e stamattina non ci fosse nulla altrimenti me ne sarei sicuramente accorto. la prima foto è una foto di ieri in cui si può vedere che lo schermo è perfetto, la seconda è terza sono di ora in cui si vedono i difetti sullo schermo? cosa posso fare? ho provato sia a fare un riavvio dei driver grafici, sia un riavvio del pc. ho letto che potrebbe essere un bug dei driver su ARM, perfavore aiutatemi. grazie mille a tutti in anticipo.
UPDATE — Amaral Turnip v2 pre-release is now available!
Thanks to the feedback received here on Reddit, we have published a new pre-release with two driver variants:
The A825 implementation is identified by its specific GPU ID, so its adjustments should not change the behavior of the other supported GPUs. This support is still experimental and needs community testing before being considered stable.
The OneUI variant is not intended to provide higher performance. It should mainly be tested by Samsung/One UI users experiencing flickering, corrupted textures, incorrect scaling, blit issues, or similar graphical problems with the standard driver.
We are also looking for feedback from users with:
Please report the device, GPU, Android/One UI version, emulator or compatibility layer, game/application, driver variant tested, and the result compared with the previous driver.
Download the v2 pre-release: https://github.com/rickamaral94/Amaral-Adreno-Tools/releases/tag/mesa-26.3.0-devel-v2
Standard driver: https://github.com/rickamaral94/Amaral-Adreno-Tools/releases/download/mesa-26.3.0-devel-v2/turnip_amaral_26.3.0-devel_v2.zip
OneUI variant: https://github.com/rickamaral94/Amaral-Adreno-Tools/releases/download/mesa-26.3.0-devel-v2/turnip_amaral_26.3.0-devel_v2_oneUI.zip
Your reports will help us validate the A825 implementation, investigate the limited improvements currently observed on A830/A840 devices, and decide which changes are safe to include in the main driver.
I bought a laptop with this chip which has 24GB of RAM. I’m mainly going to be using it for nursing school and other uses here and there. I also wanted to download Cisco Packet tracer but it said it can’t run that. I also do real estate photography and videography. For photography editing, I was able to download Lightroom Classic and use that. I haven’t done much editing yet. But I just wanted to ask what has been your guys experience with editing in this if you have this particular laptop. I don’t know what software would run well for videography editing in this, so any suggestions would be appreciated!
Thank you in advance
Hi everyone,
after sharing Snapdragon AI Studio here, I received a lot of useful feedback and questions. Thank you — several suggestions from this community have directly influenced the upcoming RC2 release.
What’s changing from RC1 to RC2:
• ControlNet Canny has been improved and integrated more cleanly into the generation workflow.
• Presets & Templates now preserve the currently selected AI model instead of unexpectedly switching models.
• Presets also preserve the active ControlNet state and reference image.
• Phoenix Boost remains available for local prompt optimization without changing the selected model.
• The Model Manager and QAI model installation workflow are being improved.
• Both Dark Mode and Light Mode are fully supported.
• The built-in user manual has been completely revised in English, German and Spanish, with additional explanations for beginners.
• Generation timing is being improved so the complete processing time can be shown, including post-processing/upscaling.
• Further work is planned around image quality, higher resolutions and additional models.
A major focus remains native local AI on Windows 11 ARM64 and Snapdragon X Elite/X Plus, especially using Qualcomm QNN / Hexagon NPU acceleration.
Some ideas raised in the discussion here — including additional models, further NPU optimization and better documentation — are also part of the roadmap.
RC2 is still under development. I’d rather release a stable update than rush the next version.
Snapdragon AI Studio is an independent open-source project and is not an official Qualcomm product.
GitHub:
https://github.com/Kreuzhofen/snapdragon-ai-studio
Which RC2 improvement matters most to you, and what would you like to see next for local AI on Snapdragon?
Thanks again for the feedback and support!
Hi everyone,
I’ve been working on Snapdragon AI Studio, an open-source desktop project exploring local AI image generation on Windows 11 ARM64, specifically for Snapdragon X Elite and X Plus systems.
Snapdragon AI Studio is an independent open-source project and is not an official Qualcomm product.
The project uses the Phoenix Engine and currently focuses on ONNX Runtime, Stable Diffusion / ControlNet workflows, and Qualcomm QNN / NPU development.
RC1 is publicly available, while I’m currently working toward RC2 with a focus on image quality, performance and further native Snapdragon acceleration.
I’d especially be interested in feedback from other Snapdragon developers and users: What local AI workloads are you currently running on Snapdragon X systems, and what would you most like to see improved for native Windows on ARM AI applications?
GitHub:
https://github.com/Kreuzhofen/snapdragon-ai-studio
— Holger Kreuzhofen
Founder & Lead Developer, Snapdragon AI Studio / Phoenix Engine