Nuvia: don’t hold your breath

3DMark can be used to compare gaming

Does qualcomm 750 gpu have mesh shaders?

No idea if the 750 GPU does or doesn’t.

The only reference I can find is Wikipedia (without a source link) that says that it does. But even if it does, is it M1 style support (emulated) or Hardware support (A17/M3)?…Very little info on the 750.
 
Last edited:
There are a few game engines that could probably be used to compare different GPU configurations running essentially identical software on Metal and Vulkan and, what, DirectX is it? Should be possible to make equivalent tests.
 
Pretty sure thats only for CPUs.

It's same problems, same data, same algorithms. Yes, you can compare Vulkan, OpenCL, and Metal. Of course, you'd be comparing not only the GPU but the tooling and the driver, but that's how things are when comparing GPUs. If it helps, you can reframe this as comparing how good are various platforms at solving the same problem. Note that the comparison problem already exists between the CPUs as well, as there are differences in the compiler, OS scheduler, memory allocator etc.
 
Anandtech look at the Elite X

interesting stuff. Linux score much higher than Windows on Geekbench single core, at least partly due to the fact that they don’t have fan control working under Linux and thus fans run without limits. Lol

Windows score: 2971
Linux: 3236

More later but there are no power figures in the article.
 
I doubt that they had the opportunity to do detailed power measurements. That said, if the 23W TDP is accurate, the 12-core performance in CB 2024 looks very nice (although I suspect the actual power consumption is higher). The lack of improvement with going to the 80W mode is alarming.

The 3000 GB6 single-core in Windows is kind of underwhelming as well, especially with the 4.3Ghz boost. I hope they will work with Microsoft to improve this before the official release.
 
Honestly, I would like to see Nuvia, et al flourish somewhat. I dislike MS Widows a lot, but nowhere near as much as I detest x86. I want to see x86 become more marginalized than, say, Amiga. They can make all the BS claims they want, but the proof will be in the custard. When 90% of non-Mac personal computers are running on ARM, people will be able to see, straight up, that Macs offer a better experience, and no slideshow in the world will be able to change that.

1698685321659.png
 
Anandtech look at the Elite X

interesting stuff. Linux score much higher than Windows on Geekbench single core, at least partly due to the fact that they don’t have fan control working under Linux and thus fans run without limits. Lol

Windows score: 2971
Linux: 3236

More later but there are no power figures in the article.
At least for the Linux high performance multicore GB6.2 benchmark we have something approaching more reasonable in terms of scores. Unsure how much power it draws to get that 17K score of course.
 
Have to say, I’m feeling pretty optimistic about Oryon.

There’s clearly room for improvement (efficiency seems questionable at this early stage, there are no E-cores, GPU performance is limited etc.), but it looks impressive for a first generation product.

Two things I’d like to see:
  1. SystemReady compliance/certification.
  2. A firm commitment to support Linux. Linux support should be a priority and at least match what Intel/AMD provides. This hardware is no good if it can only run Windows 😅
 
Have to say, I’m feeling pretty optimistic about Oryon.

There’s clearly room for improvement (efficiency seems questionable at this early stage, there are no E-cores, GPU performance is limited etc.), but it looks impressive for a first generation product.

Two things I’d like to see:
  1. SystemReady compliance/certification.
  2. A firm commitment to support Linux. Linux support should be a priority and at least match what Intel/AMD provides. This hardware is no good if it can only run Windows 😅
I’m sure it will support Linux. That’s one of the reasons they quote the geekbench figures for it.

I’m a little suspicious much of the momentum will evaporate by the middle of next year. I think if it was released this year, it would have been great. Next year less so, but we’ll see.
 
I’m sure it will support Linux. That’s one of the reasons they quote the geekbench figures for it.

I’m a little suspicious much of the momentum will evaporate by the middle of next year. I think if it was released this year, it would have been great. Next year less so, but we’ll see.
A lot depends on releases from AMD/Intel - less so on Apple. I know they compare a lot to the M2 but they’re in less direct competition with Apple than the former two. They need to get OEM buy-in for overall compelling products and not just chip specs. That’s another reason they’re trying build hype this far out. Agreed that it would seemingly be a great product if it had been announced months ago and releasing right now though.
 
A lot depends on releases from AMD/Intel - less so on Apple. I know they compare a lot to the M2 but they’re in less direct competition with Apple than the former two. They need to get OEM buy-in for overall compelling products and not just chip specs. That’s another reason they’re trying build hype this far out. Agreed that it would seemingly be a great product if it had been announced months ago and releasing right now though.
Oh agreed. There are many good reasons why they needed to announce now. I just wonder if they can maintain the good feelings this far out.
 
This thread started in late 2021 and is entitled “don’t hold your breath.”

Less talk, more shipping stuff.
It looks like they have, at the very least, engineering samples (I believe the press has been able to try some laptops with the new chip, albeit briefly). How early before release does this kind of thing happen? Time from tape out to release, I mean.
 
It looks like they have, at the very least, engineering samples (I believe the press has been able to try some laptops with the new chip, albeit briefly). How early before release does this kind of thing happen? Time from tape out to release, I mean.

“It depends?”

The x704 taped out, and when it came back we found out that we had reduced power consumption too much and caused cross-coupling problems. Other times stuff tapes out and you find bugs and have to re-spin a metal layer or two to fix them. If you are taping out a chip set, you need to take time to test everything together. If it’s a new packaging technology, you have to deal with that, too. Could be anywhere from 3 months to a year.
 
Looks like we have the first X Elite Geekbench score.

Edit: Actually there’s quite a few scores there now and all the scores over 3000 are on Linux where we know the fans were on full blast.
 
Last edited:
Back
Top