That is a false dichotomy. I own an ARM based machine. I did not lose the access to the x86 software ecosystem. Apple has demonstrated that it is possible to run x86 software on an ARM-based machine without significant drawbacks. If other vendors follow them in adopting x86 compatibility features, legacy software support won't be an issue.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Qualcomm Nuvia vs Apple Mx
- Thread starter crazy dave
- Start date
-
- Tags
- amd apple silicon intel nuvia qualcomm
- Sort by reaction score
That is a false dichotomy. I own an ARM based machine. I did not lose the access to the x86 software ecosystem. Apple has demonstrated that it is possible to run x86 software on an ARM-based machine without significant drawbacks. If other vendors follow them in adopting x86 compatibility features, legacy software support won't be an issue.
I dunno about that. For a server I’d do a lot of low power cores with a high performance fabric. For laptop I’d do a smaller number of high performance cores with a focus on heterogeneous secondary processors.
Perception doesn't pay the bills. And few companies jump at the chance to make solely low margin devices.
It's a chicken and egg problem. I'm sure they're already having meetings with OEMs like Dell, HP, Lenovo, etc. If they can deliver performance comparable to M1, in volume, then they may have a shot. M1 is already more powerful than a lot of business PCs. Perhaps a few larger customers would like to replace fleets of bulky, power hungry desktops with something more efficient. Ditto for the many laptops deployed. If MS has their solutions native, that might be enough to get a few orders. They don't have to go after gamers or high performance users yet, though I wouldn't be surprised if MS were thinking of a smaller, quieter Xbox in the future.
They just need to establish momentum. Having a decent WoA dev platform might help too.
They just need to establish momentum. Having a decent WoA dev platform might help too.
Cross out Dell. They are in bed with Intel.
If indeed. Bad news on that front…
Ya think?
Sampling 2022... 2023 release. So everyone will have possibly 2 product cycles to improve before these Nuvia SoC's see daylight.
Microsoft has WoA, and it can now run 64-bit apps. If someone can get something remotely comparable to M1, the performance should be adequate for most business tasks. Office is now native on both W11 and macOS. As long as Slack/Teams, and other corporate software can run acceptably, it shouldn't be a problem.
I also forgot Google and Chromebook, which already runs on ARM, as another source of potential sales.
In any case, I imagine there are a few OEMs looking at M1 and wanting something comparable to offer their customers, or at least have the option for leverage against Intel/AMD.
Well if we take Nuvia’ pre-acquisition marketing numbers at face value (I know, I know, but for the sake of argument), then they have a core with as good if not better perf/W curve than firestorm. Firestorm is obviously good enough as a laptop processor and is low power enough for a server. Therefore these theoretical cores should be good enough to fill both roles, no?
However the fabric and efficiency cores are a definite issue. The interview says Nuvia will be a completely new SOC package which at least requires new efficiency cores. Maybe they could use Arm A5x cores but licensing and engineering issues have historically meant that was substandard pairing. Maybe it would go better. No word on whether Nuvia had its own fabric that can be used or if Qualcomm’s will/can be used.
Claiming performance/Watt numbers for a processor you haven’t even taped out (did they?) takes a particular kind of chutzpah.
Absolutely no disagreement … and no they didn’t. Those were from simulations of their designs.
Good luck “simulating” power usage accurately.
Not to mention that just because you tell a simulator that your processor runs at X GHz doesn’t mean you know how to do the physical design so as to achieve that.
Had a conversation once with AMD’s former CTO about this sort of thing - we both believe you make a chip, then you report the results.
But that sounds like work … and somewhat anti-synergistic with getting invested in (or in their case bought over) before having to spend the money actually building something
Your reply has little substance for the supply chain.
People will buy it when they perceive it to be a better offer than Intel/AMD.
As ~80% of all laptops/desktops are below that price point then ARM Android SoC makers are more likely to tackle that first.
That's how Apple did their transition from Intel (100%) > M1 (~80%) > M1 Pro/Max (~20%).
Don't tell me you know better than Tim Cook.
Assuming ARM Android SoC makers execute things properly I expect them eating into ~80% of Intel/AMD CPU & SoC market share as early as 2030.
What is lacking are Mac-quality implementation of the following
- Win11 on ARM
- Rosetta-like translation layer so that legacy & abandoned software will run at near native speeds
- fat binaries that will work on both ARM & x86 for future software
I hope that Win11 on ARM will out compete Win11 on x86 on price so performance per watt improves, raw performance improves and average selling price of hardware drops further to say ~$300 by 2030s for laptop/desktops.
Dont believe me? Look at what happened to Nokia & Blackberry as early as 2017 when the iPhone X came out.
Unlike the PC industry the smartphone industry is much more cut throat and competitive.
An indicator that Intel sees this as a likely possibility is their offer to be a foundry for rival chip brands.
Last edited:
It's quite obvious that Microsoft does care about Windows on ARM. There are huge financial implications for Microsoft to care.
The main reason is that the M1 is so far ahead of what Intel and AMD can provide, people can/will switch to macOS. In order to better compete with Apple, Microsoft needs to make Windows on ARM work and work extremely well so that it opens up to more competitors such as Qualcomm, Mediatek, Ampere, etc. In fact, Microsoft itself is designing ARM chips internally for Azure and their Surface line of laptops.
Will the Windows transition be as smooth as macOS'? Nope. Apple controls the entire stack and cares way less about legacy compatibility. But I do think Microsoft will get there one day and they have the desire to.
So far Samsung’s leading nodes have been a step behind TSMC’s. But they have been improving, all it takes is a stumble by TSMC.
So its quite possible that a Nuvia chip won’t match the single-threaded performance of M1, just because they haven’t been as focussed on that aspect of performance. Interesting.
Anyway I’d expect the M2 to come in 2022, but likely the M3 not until 2024.
Unclear on all counts.
As @Bodhitree said Samsung’s leading edge node has recently been about a generation in performance and power behind TSMC’s despite using the same node name. However, Samsung is adopting GAA technology first. TSMC isn’t planning on implementing GAA until 2nm. So maybe that’ll make a difference and soon Samsung will catch up or surpass TSMC in the upcoming generation, but no one really knows.
Qualcomm has and will use any foundry they can get space for.
I expect Nuvia to match and beat the M1 in ST performance. This is a 2022/2023 chip going up against a 2020 chip. The main reason Qualcomm spent $1.4b on them is to get a high-performance core, able to compete with Apple Silicon. And since they're publicly saying that they're aiming at Apple Silicon, they must be confident that it will either match or surpass the M1. Anyhow, competition is good. I'd love to see Apple go on a 12-month cadence to upgrade the M, M Pro, and M Max chips. Maybe they can go 2 years for the Duo/Quad versions.
Don’t bother, the dude is either a shill or a really, really sad person who feels the need to take up the sword for PC/x86.
I’m not sure Microsoft is focused enough to make the jump though. Windows isn’t their only breadwinner anymore with things like Azure and Gaming (they seem to be pushing this hard as of late).
They’re not married to x86 of course, but I don’t think they’re that invested in making Windows platform agnostic either.
They are investing in making it agnostic, or as much as possible. It's not a macOS like transition. But Windows on ARM emulates x86 apps. On the Windows store, you can submit ARM and x86 versions of the app.
Microsoft isn't doing a hard transition like Apple is because they can't just cut out AMD/Intel. There's way too much legacy stuff. And no one has built a faster CPU for Windows than Intel and AMD.
By the way, Windows isn't Microsoft's only breadwinner but the Mac is pretty far down Apple's priority too.
I think it's going to take Microsoft a much longer time to do the transition. But I think Windows will fully support ARM which is good news for Apple Silicon Macs because more software will come optimized for ARM, including games.
Apple’s strategy as of late seems to be using their hardware as an entry point to their services. The Mac, as of late, I’d say has relatively high priority. But that’s aside from the argument.They are investing in making it agnostic, or as much as possible. It's not a macOS like transition. But Windows on ARM emulates x86 apps. On the Windows store, you can submit ARM and x86 versions of the app.
Microsoft isn't doing a hard transition like Apple is because they can't just cut out AMD/Intel. There's way too much legacy stuff. And no one has built a faster CPU for Windows than Intel and AMD.
By the way, Windows isn't Microsoft's only breadwinner but the Mac is pretty far down Apple's priority too.
It’s difficult to compare since Microsoft is a different beast entirely.
From what I’ve read, Microsoft has a lot of competing teams that don’t necessarily work together, and aren’t really focused on anything generally. Things like WSL and ARM support get added, but not improved much as attention shifts to another feature.
This may not necessarily be the case, it’s anecdotal, but it seems consistent with my observations.
If attention shifts back to full-fat ARM support then I could see it happening, but other than that it’s unlikely imo.
What is "full-fat ARM support"?
I think Microsoft's current level of commitment to Windows on ARM makes sense.
Well WSL2 was an improvement over WSL and so far MS is supporting and updating WoA. Full emulation/translation while late considering is here and I think most would argue that the thing holding it back is hardware not OS.