• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Discussion Qualcomm Snapdragon Thread

Page 203 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
gdansk said:
Zenbook A16 has arrived. Screen looks great. No notch. Ceraluminum might be a marketing name but it feels nice whatever it is. Bloatware and stickers. They never learn. Fan noise is tolerable in normal usage. In whisper mode I haven't heard it using the web. Maybe they finally did it. In performance mode it is comically loud, so I am avoiding that even for the tests here. Track pad isn't bad except the click mechanism but you can avoid that.

It wants a 130W charger. That sucks because everyone in my family uses Apple laptops so our homes are littered with their 100W chargers. It'll be weeks before I know typical battery life.

It really sucks in Firefox. About 1/2 the score of a 9950X3D in Speedometer 3.1. Does okay in Chromium-based browsers. ~40 compared to ~50 for 9950X3D. It is a laptop, however, and it feels responsive so far despite that.

GB6.7

ASUS Zenbook A16 UX3607OA - Geekbench

Benchmark results for an ASUS Zenbook A16 UX3607OA with a Snapdragon X2 Elite Extreme - X2E94100 - Qualcomm Oryon processor.
browser.geekbench.com browser.geekbench.com
GB6.2 (No SME)

ASUS Zenbook A16 UX3607OA - Geekbench

Benchmark results for an ASUS Zenbook A16 UX3607OA with a Snapdragon X2 Elite Extreme - X2E94100 - Qualcomm Oryon processor.
browser.geekbench.com browser.geekbench.com
GB5.5

ASUS Zenbook A16 UX3607OA - Geekbench

Benchmark results for an ASUS Zenbook A16 UX3607OA with a Snapdragon X2 Elite Extreme - X2E94100 - Qualcomm Oryon processor.
browser.geekbench.com browser.geekbench.com

Cinebench 2026
Normal performance mode
616 ST
5977 MT
9.71x

Benchlight 0.9.7 (WSL)
101.gcc: 11.05
102.xz: 3.66
Click to expand...
Nice. Can you use hwinfo to check power draw in ST results? Allegedly Qcomm added support for this, in this gen.
1775764544996.png
 
CouncilorIrissa said:
Not bad. Not M5 (spaceship) territory, but quite a bit ahead of the Windows competition.
Click to expand...
My initial impressions keep getting better and better.
  1. Driver software for my Logitech mouse just worked.
  2. I'm compiling projects now and it is faster than any other laptop I've used that can run full Visual Studio. The fans didn't turn on in the time it took to build.
  3. Some old 32-bit only game I play works in emulation. I really did not expect that. It barely works on modern Windows x86.
  4. WSL just worked and installed aarch64 Linux.
The biggest negative impressions so far have all been Asus/Microsoft bloat and shenanigans. Those are outside of Qualcomm's control.

Geddagod said:
Nice. Can you use hwinfo to check power draw in ST results? Allegedly Qcomm added support for this, in this gen.
...
Click to expand...
I tried running the beta version. I'm not sure why if it's 1T it has two clusters active.
1775764900831.png
 
gdansk said:
Depends on the mode. Whisper mode reports ~50W. Standard mode ~75W. Performance mode ~100W.
I'm not sure how accurate this is. And my bet is that vendors can play stupid power limit games given these are Asus power modes.
Click to expand...
You can measure at the wall. But yeah looks about right for a cpu with 18 cores on N3X.
 
gdansk said:
My initial impressions keep getting better and better.
  1. Driver software for my Logitech mouse just worked.
  2. I'm compiling projects now and it is faster than any other laptop I've used that can run full Visual Studio. The fans didn't turn on in the time it took to build.
  3. Some old 32-bit only game I play works in emulation. I really did not expect that. It barely works on modern Windows x86.
  4. WSL just worked and installed aarch64 Linux.
The biggest negative impressions so far have all been Asus/Microsoft bloat and shenanigans. Those are outside of Qualcomm's control.


I tried running the beta version. I'm not sure why if it's 1T it has two clusters active.
View attachment 141523
Click to expand...
Wow yeah, that's super weird. Are the other cores also boosting in the other clusters?
 
gdansk said:
My initial impressions keep getting better and better.
  1. Driver software for my Logitech mouse just worked.
  2. I'm compiling projects now and it is faster than any other laptop I've used that can run full Visual Studio. The fans didn't turn on in the time it took to build.
  3. Some old 32-bit only game I play works in emulation. I really did not expect that. It barely works on modern Windows x86.
  4. WSL just worked and installed aarch64 Linux.
The biggest negative impressions so far have all been Asus/Microsoft bloat and shenanigans. Those are outside of Qualcomm's control.


I tried running the beta version. I'm not sure why if it's 1T it has two clusters active.
View attachment 141523
Click to expand...

Any chance you feel like running some other benchmarks? Perhaps source-based benches that imitate SPEC subtests?
 
gdansk said:
My initial impressions keep getting better and better.
  1. Driver software for my Logitech mouse just worked.
  2. I'm compiling projects now and it is faster than any other laptop I've used that can run full Visual Studio. The fans didn't turn on in the time it took to build.
  3. Some old 32-bit only game I play works in emulation. I really did not expect that. It barely works on modern Windows x86.
  4. WSL just worked and installed aarch64 Linux.
The biggest negative impressions so far have all been Asus/Microsoft bloat and shenanigans. Those are outside of Qualcomm's control.


I tried running the beta version. I'm not sure why if it's 1T it has two clusters active.
View attachment 141523
Click to expand...
It sounds like ARM Windows is maturing up nicely.
 
gdansk said:
My initial impressions keep getting better and better.
  1. Driver software for my Logitech mouse just worked.
  2. I'm compiling projects now and it is faster than any other laptop I've used that can run full Visual Studio. The fans didn't turn on in the time it took to build.
  3. Some old 32-bit only game I play works in emulation. I really did not expect that. It barely works on modern Windows x86.
  4. WSL just worked and installed aarch64 Linux.
The biggest negative impressions so far have all been Asus/Microsoft bloat and shenanigans. Those are outside of Qualcomm's control.


I tried running the beta version. I'm not sure why if it's 1T it has two clusters active.
View attachment 141523
Click to expand...
Can you try running some games with Prism or FEX?
 
Nothingness said:
Isn't FEX Linux-only?
Click to expand...
According to this, you could use it as the emulation back end in Windows too...


This is how the OS knows where to locate a 32-bit emulator (whether hardware sandbox or software translator) to inject into the 32-bit process. You should now realize that the x86-to-ARM64 translator is a pluggable component listed in the registry. This is the key to plugging in other implementations from Microsoft or third parties. The Wine 9.0 release notes recently disclosed that Wine now implements that same registry key, opening the door for third party emulators on Linux such as FEX to plug in. Very cool! Obviously the emulator that kicks in on macOs is Rosetta today. What would be far more interesting to me (as a Windows user and developer) would be to see a third party emulator used for Wine on Linux get ported over to Windows as an alternative to xtajit.dll and xtajit64.dll.

In the case of 64-bit emulation, there already is a pluggable third party emulator (of sorts) - the xtabase x64 interpreter which I'd written as a test and validation tool has been included in both Windows 11 SV1 and SV2 (builds 22000 and up) for almost three years. It was my original predecessor to xtajit64.dll during development and features extras such as: implementation of the CPUID ABM feature (which is mostly just implementing the LZCNT instruction which AMD had introduced over 15 years ago), very accurate 80-bit x87 floating point emulation using a pure software 80-bit floating point emulation package, and more accurate and deterministic EFLAGS arithmetic flags emulation than JIT translations. That is why xtabase was used as the reference interpreter when getting the JIT up and running.
Click to expand...
 
Tachyonism said:
Can you try running some games with Prism or FEX?
Click to expand...
I don't have benchmark results but I can say the following games worked or not:
1. Forza Horizon 4 - crashed to desktop
2. Diablo 2 Resurrected - playable
3. Civilization V - playable
4. Age of Empires 2 DE - playable

Not a really recent collection but I was trying games I thought I might actually play on my laptop instead of desktop.
 
trivik12 said:
@gdansk how is the performance on battery? is it noticeably slower?
Click to expand...
It wasn't noticeably slower. However, today I am testing it docked with 4K 120hz monitors.
Also I have extracted the ACPI Evaluate Output Buffers. Might be useful for future Linux support, can't hurt to have them public anyway.
 
Last edited:
Are there any rumors on when the next big Oryon arch update is supposed to come? I don't follow ARM leaks as much as I do for x86 stuff.
For the past 2 gens Qcomm has gotten very little in terms of IPC uplifts and from a block diagram it doesn't look like Qcomm is doing much other than rebalancing (not even growing) some structures.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top