Qualcomm revealed X Elite's benchmark scores

[MayaUser]

Did Qualcomm mention whether Orion is ARMv9 or ARMv8? I can't seem to find this info.

They run the Arm v8 instruction set, specifically, Arm v8.7
Each cluster of four CPU cores has 12MB of level-2 cache attached
There are three reference platforms being shown off are 12W, 23W, and 45W, with the latter able to scale to 80W if necessary

 

[leman]

They run the Arm v8 instruction set, specifically, Arm v8.7
Each cluster of four CPU cores has 12MB of level-2 cache attached
There are three reference platforms being shown off are 12W, 23W, and 45W, with the latter able to scale to 80W if necessary

This really sounds more and more like a copy of Firestorm… did Williams and his team simply recreate what they did at Apple? That’s a bit disappointig…

 

[APCX]

They run the Arm v8 instruction set, specifically, Arm v8.7
Each cluster of four CPU cores has 12MB of level-2 cache attached
There are three reference platforms being shown off are 12W, 23W, and 45W, with the latter able to scale to 80W if necessary

Huh…interesting. Do you happen to know if the benchmarks shown mix these platforms? So the same performance as M2 at 30% less power might be one platform, but the 2x performance compared to M2 might be the 45W platform.

I thought there was just one platform tbh.

 

[MayaUser]

This really sounds more and more like a copy of Firestorm… did Williams and his team simply recreate what they did at Apple? That’s a bit disappointig…

Yes, and it just shows that Johny Srouji is still the brain of chip architecture, his people if they will leave they will just copy and paste in most part what they did in the past. This happened when he was at Intel. As long as he is at Apple, this company will still have good SoC

More about this will come in Spring, preparing and some demo from the 9 companies that will adopt this :
Acer
Asus
Dell
HP
Honor
Lenovo
Microsoft Surface
Samsung
Xiaomi
Samsung intend to demo in early 2024 with an release date for the public after March with others starting with late May
I personally suspect when this will be adopted by all those OEM and sold to the consumers, Apple will prepare to present the M4 with support for TB5

 

[Xiao_Xi]

This really sounds more and more like a copy of Firestorm… did Williams and his team simply recreate what they did at Apple?

How can Oryon be a copy of Firestorm if Firestorm is only v8.5 compatible? Unless A17 is v8.7 compatible, no other Apple SoC has implemented as many extensions as Oryon.

 

[deconstruct60]

Did Qualcomm mention whether Orion is ARMv9 or ARMv8? I can't seem to find this info.

" ... Unfortunately, we don’t know much more about Qualcomm’s Snapdragon X Elite chips. They run the Arm v8 instruction set, specifically, Arm v8.7, Nitin Kumar, senior director of product management said. Each cluster of four CPU cores has 12MB of level-2 cache attached, a detail that matters to chip nerds. ..."

Qualcomm's new Snapdragon is the chip we begged Intel to make

Qualcomm's upcoming Oryon CPU inside the Snapdragon X Elite is entirely made up of performance cores for maximum PC performance. Why won't Intel follow suit?

www.pcworld.com

[ I haven't second sourced that v8.7 reference. I'm pretty sure I read it somewhere but can't find it at the moment. And not sure it appeared in the video sessions. ]

Remember, the Arm version number would have been selected by limited budget Nuvia more than several years ago.

However, 8.0, 8.1 , 8.2 are not 8.7 . There is a substantive "new stuff" intersection between 8.7 the early 9's. 8.7 leaves a few in the 'optional' as a requirement but that doesn't mean they have to not be there. Just that they are not mandated.


P.S. For context. Neoverse V1 is 8.4 . So past that. Neoversion V2, N2 are 9.0. A15/A16 are repordedly v8.6 ( so would be past Apple).


P.P.S. 8.7A is paired up with 9.3A

" ...

Armv8.7-A and Armv9.2-A​

• Enhanced support for PCIe hot plug (AArch64)
• Atomic 64-byte load and stores to accelerators (AArch64)
• Wait For Instruction (WFI) and Wait For Event (WFE) with timeout (AArch64)
• Branch-Record recording (Armv9.2 only)

..."

Documentation – Arm Developer

developer.arm.com

 

[leman]

How can Oryon be a copy of Firestorm if Firestorm is only v8.5 compatible? Unless A17 is v8.7 compatible, no other Apple SoC has implemented as many extensions as Oryon.

8.7 are just a few incremental features. What I mean is that on the surface, there doesn’t seem to be much difference between Oryon and Firestorm/Avalanche (I lump them together because these are just minor tweaks). Same IPC, roughly same feature set (no big new stuff like SVE), same cluster organization…

 

[deconstruct60]

This really sounds more and more like a copy of Firestorm… did Williams and his team simply recreate what they did at Apple? That’s a bit disappointig…

Like nobody else is implementing > v8.5 ... It isn't necessarily just a copy of firestorm . Arm has v9 cores and they shared data with Nuvia to help them as a start up. ( part of reason why Arm is suing Qualcomm. ). the meme that nobody but Apple can implement leading edge Arm versions is a little nuts.

 

[leman]

Like nobody else is implementing > v8.5 ... It isn't necessarily just a copy of firestorm . Arm has v9 cores and they shared data with Nuvia to help them as a start up. ( part of reason why Arm is suing Qualcomm. ). the meme that nobody but Apple can implement leading edge Arm versions is a little nuts.

That’s not at all what I am saying. ARM CPU IP is well ahead of Apple cores in version support for example. I just find it very striking that the Nuvia core is seemingly so similar to Apple cores - down to basic organization, features (or lack there of), and even IPC!

 

[MRMSFC]

This really sounds more and more like a copy of Firestorm… did Williams and his team simply recreate what they did at Apple? That’s a bit disappointig…

If you can’t beat ‘em, join ‘em.

 

[Homy]

So products with SnapDragon X Elite will hit the market by the time we get A18 and are getting ready for M4? People say Apple is in BIG trouble or even doomed but the more important question is what would happen next if SnapDragon X Elite is in reality ”better” than Apple Silicon? Would that be a world changer? Would Mac users ditch all their Macs and go buy ARM PCs? Would Apple go bankrupt? The answer is no. Because ARM PCs don’t run macOS and using Macs, iPhones or iPads is more than just having the fastest CPU or GPU. Neither is Apple always competing with everyone. They compete with themselves by doing their best and what’s best for their users.

I also agree with Ars Technica saying ”No matter how good the hardware is, it's only part of the puzzle. The Apple Silicon transition succeeded in part because Apple's Rosetta 2 compatibility layer made x86-to-Arm code translation mostly invisible to the user, and because the company had already broken legacy technologies like 32-bit app support and certain kinds of drivers in previous macOS releases. Qualcomm will be relying on another company's software to succeed, namely the Arm version of Microsoft's Windows 11. Sometimes, even apps with Arm-native versions will download and install their x86 versions by default on the Arm version of Windows, leaving performance on the table. Drivers written for the x86 versions of Windows, and the hardware that relies on those drivers, won't work. Many games, particularly those that depend on non-DirectX 12 APIs and anti-cheat software, won't run. Backward-compatibility is one of the defining features of the x86 versions of Windows; it's much more limited in the Arm version".

 

[APCX]

Does anyone know if the X Elite uses UMA? Does it have a more traditional gpu memory model? Also does it use an TBDR?

 

[leman]

Does anyone know if the X Elite uses UMA?

Yes, just like any other SoC made in the last decade.

Does it have a more traditional gpu memory model?

The memory is shared between all the SoC processors. But it will also likely support third party GPUs over PCI-E

Also does it use an TBDR?

The GPU is a Qualcomm Adreno which is TBR (without D)

 

[jdb8167]

People say Apple is in BIG trouble or even doomed but the more important question is what would happen next if SnapDragon X Elite is in reality ”better” than Apple Silicon? Would that be a world changer? Would Mac users ditch all their Macs and go buy ARM PCs?

I might switch to an X Elite laptop depending on how well it runs Linux. I’m admittedly not a typical Mac user though—I spend most of my day in text editors and the terminal. I would never switch to Windows though.

 

[dgdosen]

Kudos to Qualcomm... Ship 'em! This also looks to be a great motivator for Apple. Maybe this is why we're seeing M3 tomorrow?

 

[dgdosen]

This really sounds more and more like a copy of Firestorm… did Williams and his team simply recreate what they did at Apple? That’s a bit disappointig…

Any insight as to why/how it looks to beat the M1 family handily?

 

[donth8]

Any insight as to why/how it looks to beat the M1 family handily?

They are boosting the clocks up to 4.3ghz for a couple cores and are only using 12 performance cores with no efficiency ones. Lastly they are ok with using 80W of power which is even higher than the M2 Ultra.

 

[leman]

Any insight as to why/how it looks to beat the M1 family handily?

Higher frequency. Not to mention they only show Oryon being faster than M2 under turbo (4.3ghz) operation, “normal” operating frequency should be much closer to M2 Pro/Max. And in MC they only claim “50% faster” than M2, which is again in line with M2 Pro/Max and a bit on a low side for 12 performance cores.

To me this looks pretty much like a rebuilt version of Firestorm/Avalanche, obviously with a slightly different feature set, tweaked differently, and manufactured on a smaller node.

 

[leman]

Maybe this is why we're seeing M3 tomorrow?

Doubt it. It’s not like they can accelerate production on a whim anyway. At any rate, M3 series were always scheduled to ship long before first Oryon laptop will be commercially available, so it doesn’t affect Apple much.

But it’s possible that we are seng the opposite: Qualcomm racing to get the press release out before M3 is announced. Then again, I doubt it. Qualcomm has good reasons to do an early announcement, they have to generate enough hype to get the ecosystem ready. That’s also how Apple dud the ARM transition.

 

[Xiao_Xi]

Qualcomm racing to get the press release out before M3 is announced.

I doubt Qualcomm knew that Apple would announce the M3 a week after the Snapdragon Summit in June.

Cristiano Amon, the president, and CEO of Qualcomm, discussed the Oryon chips in a recent investor call and specified that the company will announce specs for the processors at Snapdragon Summit 2023. That event runs from October 24-26, 2023, so we'll still have to wait a few months to learn specifics.

Qualcomm Oryon chip specs to be unveiled at Snapdragon Summit 2023

Qualcomm has high expectations for its upcoming Oryon processors.

www.windowscentral.com

 

[APCX]

I doubt Qualcomm knew that Apple would announce the M3 a week after the Snapdragon Summit in June.

Qualcomm Oryon chip specs to be unveiled at Snapdragon Summit 2023

Qualcomm has high expectations for its upcoming Oryon processors.

www.windowscentral.com

They wouldn’t know the exact date, but they knew it would arrive before their own effort.

 

[Xiao_Xi]

they knew it would arrive before their own effort

I don't know how. Up until last week, rumors were that Apple wouldn't launch the M3 until spring 2024.

 

[leman]

I don't know how. Up until last week, rumors were that Apple wouldn't launch the M3 until spring 2024.

And these rumors might well be right. It’s possible we will only see the Pro/Max laptops tomorrow, not the base M3.

 

[APCX]

I don't know how. Up until last week, rumors were that Apple wouldn't launch the M3 until spring 2024.

And they know theirs won’t arrive until mid-2024.

 

[DeepIn2U]

A lot of hate going round, versus being happy there is finally some decent competition - even if its shortly to be 1 generation behind on performance. Apple Mac users SHOULD be Happy.

It says "matching" uses 30% less power - not beating it.

^ THAT is the key part you're missing. Everything else you've mentioned may be valid but the KEY part is MATCHING with 30% less power - this is significant if it holds proven on a shipping product Summer of 2024.

This is the competition that'll push Apple's engineering team further. The industry and consumers will benefit and rejoice.

Also Cannon's finding of an existing way to mac SoC's that previously proved difficult has been resolved.
Nanoimprint lithography semiconductor manufacturing system that covers diverse applications with simple patterning mechanism.

Also interesting to note that it's not a big.Little design. It has 12 homogenous cores and 2 of them can boost to 4.2Ghz.

I think its efficiency will suffer compared to Apple Silicon on Macs.

It does make sense though. Nuvia was designing a server core before the acquisition. There was no reason to design 2 separate cores for servers.

This is an SoC based on ARM which uses RISC ... it IS a big.Little design, period. How the cores are set in terms of performance does not change this.