5 ms·
It's funny, I honestly doubt Qualcomm can compete at this point. Intel intends to be on the same node as Apple by EOY 23, so honestly Qualcomm would have to mak
by smoldesu 5y ago
It's funny, I honestly doubt Qualcomm can compete at this point. Intel intends to be on the same node as Apple by EOY 23, so honestly Qualcomm would have to make a pretty competitive chip to have a chance at playing in the big leagues. Look at it this way: Qualcomm needs to have CPU performance that can beat Intel and iGPU performance that can compete with Ryzen chips, that's the entry-level requirement to start convincing people that ARM is the future on PC. Of course, this is not going to happen, especially if they're on some podunk 7nm Samsung lithography or something of that vain.
I can see this unfolding one of two ways, neither of which are good for Qualcomm:
1. Intel takes back the performance crown and recoups their efficiency losses over the next 2 or 3 generations. They double or triple their density, start aggressively buying up TSCM silicon to supplement their own fabs, and go back to being the de-facto PC CPU manufacturer.
2. AMD pushes hard on their efficiency, 3D cache and iGPU technologies, muscling Intel out of the competition and making an indisputable case for their dominance in the mobile/laptop category. Everyone except long-time Intel contractors switch their main products for Ryzen chips, and eventually they get the leverage they need to take over the professional/business sector too.
Notice how nowhere I said "Qualcomm, a company with little to no history in the field of PC manufacturing, designs a chip that will likely be slower than it's x86 counterparts, has no means of emulating the x86 software people buy Windows machines for, and magically finds it's way into hundreds of consumer machines." That's because it's a pipe dream, and with the nature of PC computing, you simply can't transition architectures in the same way Apple did. I reckon x86 is going to stay relevant for at least another 5-10 years on the Windows/Linux side of things, and by the time people are sick and tired of it, I think we'll be eyeing RISC-V chips by then. ARM is starting to get long in the tooth, and it's advantages as an architecture have pretty substantially diminished over the last 20 years. It also still hasn't solved big problems like SIMD execution or ISA modularity yet; there's a real chance for another architecture to eat it's lunch in the distant future.
- aDfbrtVt 5y agoWhy would Qualcomm design their silicon for an 8nm Samsung process? They have a working relationship with TSMC and definitely have enough volume to access risk production nodes.
- singhrac 5y agoI don't have extremely high confidence in Qualcomm as a company either, but there are many different paths to Qualcomm becoming competitive, with the right management of the Nuvia team. They can (and should) manufacture on the Intel (TDM 2) or TSMC nodes to stay competitive, and Microsoft is clearly willing to partner on the software side to make the x86-to-ARM transition possible. I suspect Samsung might actually get there first since they are more similar to Apple and have long produced good (if not top-of-class) phone chips and are integrating AMD's RDNA2 into their Exynos chips. Still, it's clear Apple is a good couple of years ahead, I don't expect 2023 Qualcomm or Samsung platforms to be competitive with M1, let alone M2.
- Sakos 5y agoSamsung is irrelevant in this discussion as long as their process/node issues aren't resolved and there's no indication that they're anywhere close to doing so. Previous performance doesn't matter if they're unable to make new, competitive nodes work.
- klelatti 5y ago> ARM is starting to get long in the tooth That’s Arm v8 that was announced in 2011?
- Arcuru 5y agoI have no argument with most of that, but Qualcomm and Microsoft have been shipping ARM devices with x86 emulation for 4/5 years now.
- Melatonic 5y agoI'm wondering if we are eventually going to get hybrid chips that have X86 big cores and ARM "little" efficiency cores. Is it not possible to have an OS that can run services on both simultaneously?
- wmf 5y agoThis would be nothing but pain. If you want efficiency cores you can use x86 for that.
- smoldesu 5y agoNot really, at least with how it currently works. Most OSes will redirect threads to the different cores based on their resource usage, which is impossible if each core was running a different architecture. No amount of magic will make it work elegantly, I'm afraid.
- nicoburns 5y agoWhat would be the advantage? Apple have shown that ARM cores can compete at the high power end, and Intel's little cores aren't that far off at the low power end.
- Const-me 5y agoWhy do you think the market inevitably becomes a monopoly? In the last 20 years we had generations when AMD CPUs were better (Athlon XP, Athlon 64, Zen 2, Zen 3) and generations when Intel CPUs were better (Core 2 Duo, Sandy Bridge, Haswell, Skylake). That competition pushed performance to the awesome levels we have now, when high-end desktop CPUs exceed 1 TFlops FP32. As a consumer, I find the trend awesome and I hope it continues in the future.
- andjd 5y agoCounterpoint: X86 is dying. It is my understanding that Intel/AMD will have a hard time making x86 chips have a competitive performance-per-watt with ARM chips. This is because x86 has a large instruction set, and those instructions must be physically implemented in the silicon. It's why you never had an x86-powered smartphone. Efficiency is a killer app for laptops. It lets manufactures design thinner lighter computers with all-day or longer battery life. This is more valuable to most consumers than support for every legacy app that doesn't get ported to run natively on ARM. Most recent software written in C#, Java, or Electron can be trivially recompiled to target ARM architectures. There's a chicken-and-egg situation here, but Apple's M1 launch is likely to push Microsoft off the fence and get them to seriously support Windows-for-arm. Microsoft's x86 emulation is currently bad, but if they can make it half-decent, the corpus of software that can't or won't be ARM-native will be usable -- if slow -- so this won't substantially impair adoption of ARM PCs. There will still be some niche users who need run software that's optimized for x86, in the same way there are niche users who still need to use parallel ports or PCI cards. The industry will support those users, but because they are a niche, volumes will be low and prices high. Also important here, is that besides competing with Macs, Windows PCs are also competing with Chromebooks. Chromebooks have already dominated the large education market, and companies are beginning to make premium Chromebooks because there's a whole generation of young adults that grew up using them and it's the computer platform they're most comfortable with. There are already ARM-based Chromebooks, so this is a platform where backwards compatibility with x86 is largely a non-issue. This will further push Microsoft to support ARM chips to avoid becoming an also ran. I'm not saying this is how it's going to be. But your take is pretty maximalist on x86. There's a similar story where x86 becomes basically irrelevant in 5 years or less.
- smoldesu 5y agoI understand where you're coming from, but I don't think you have the entire story; x86 is old and massive. You're correct, it has hundreds upon hundreds of instructions, many of which require hardware accommodations that can be fairly expensive to implement on-die. On top of that, it has a bevvy of ISA extensions like SSE and AVX that require even more hardware, and can drive complexity even further. The issue is that ARM is getting to the same point. ARM is no longer RISC; it's instruction set is anything but reduced. Much like x86, it has hardware-specific instructions that correlate to esoteric software features like Javascript execution or OS sandboxing. It also has lots of extensions like NEON, as well as Apple-specific ones that cannot be open sourced or replicated by third parties that want to also adopt ARM. The push to "modularize" ARM has failed, so much so that we update the spec with new versions requiring specific extensions to qualify. ARM from a hardware implementation perspective has become a mess, arguably moreso than x86 has. On top of that, it's extensions are fairly disappointing when compared to x86 ones. NEON gets smoked by AVX by 2 orders of magnitude. Floating point math is utterly lethargic without hardware extensions. Extended precision on ARM is simply not very well-supported. All of these things come back to bite it when considering as a desktop architecture, and ultimately represent a conflict of interests with it's application; ARM was split between running on $1 microcontrollers and $1,000 iPhones. There was no good solution that would please everyone, so they tacked on complexity until both sides were happy. And that leads us to where we are today, with two bloated architectures fighting it out for superiority on completely separate terms. x86 is by no means a perfect architecture, but ARM also suffers a lot of the same issues. They're both old, they're both extremely specific in implementation and require lots of licensing to produce. The advantages of switching to ARM are controversial and marginal, much like the case for moving away from x86. On top of that, it's not like 5 watt x86 chips don't exist; they simply aren't comparable to the M1 or other high-performance ARM chips because they're manufactured on silicon that's 2-3x less dense. On paper, x86 has a lot of efficiency it can recoup by moving to more recent lithography, and until Apple is fighting x86 manufacturers on equal footing (which should kick up around EOY '23), I think it's too early to say who's "won" the desktop market. We simply don't know, but counting x86 out of the game is a pretty hype-driven musing in my opinion.
- faeriechangling 5y agoI'm far more bullish on Qualcomm here. I just think they're in a great strategic long term position. The writing is on the wall when it comes to x86 for mobility, x86 is crap for mobility, ARM is fast enough to emulate legacy x86 apps, so for me the question isn't "Is ARM viable" it's "Can Qualcomm beat AMD/Intel's/Apple's non-ARM architectures"? In that respect, I see Qualcomm's strategic positioning as being great because they power Android. I can only see them coming after Intel's Celeron line with a vengeance in the medium term, and I don't think they really be more than in the ballpark when it comes to performance, I think Qualcomm needs to compete on cost and power efficiency. I don't really see RISC-V being a thing any time soon because smartphones run ARM, they don't run RISC-V.
- smoldesu 5y agoI don't really get your argument here, hopefully you can clear some stuff up for me. > The writing is on the wall when it comes to x86 for mobility, x86 is crap for mobility Is it? Or are legacy manufacturers just bad at staying up-to-date with the latest node advancements. The only reason Apple has made a consistent case for ARM is because they're using lithography that's more than twice as dense as Intel and ~50% more dense than AMD. When you map out their power efficiency, I think most of it tracks. If Apple built the M1 on x86, I'd wager it would have the same battery life, +/-10%. > ARM is fast enough to emulate legacy x86 apps Nitpicking, but Rosetta isn't an emulator so much as it is a dynamic recompiler. The code you're running is translated, not interpreted, which is why it can even run at near-native speeds in the first place. > In that respect, I see Qualcomm's strategic positioning as being great because they power Android. I can only see them coming after Intel's Celeron line with a vengeance in the medium term, and I don't think they really be more than in the ballpark when it comes to performance, I think Qualcomm needs to compete on cost and power efficiency. This much I can understand, and I think you're spot on with how they're going to compete with the Celeron line. Now that Alder lake is rolling out though, I think things are going to get really heated. Intel can now compete on power efficiency, and it's only going to get better as they focus on density and pushing heterogeneous architectures. Qualcomm is going to have a hard time competing with TSMC 5nm 8c16t Ryzen chips and 4e/4p big.LITTLE Intel chips, since both of them should have pretty good power efficiency. That leaves you with cost, which I reckon AMD will struggle with, but Intel should be able to price pretty competitively courtesy of their more mature silicon. Time will tell, I guess. > I don't really see RISC-V being a thing any time soon because smartphones run ARM, they don't run RISC-V. I mean, yeah. RISC-V didn't even exist when ARM was being considered for smartphones, but that hasn't held it back. Android runs just fine on RISC-V today, the big reason it's not being used is because there are no major manufacturers who produce RISC-V chips, leaving them fairly expensive and out-of-reach for phone companies. That price is expected to nosedive over the next few years though, especially with the insane level of investment from countries like China. Since there's no licensing involved with making RISC-V chips, manufacturers will have to compete directly on performance and pricing merits alone. It's whataboutism to be sure, but things are looking really good for RISC-V usurping ARM in the next 5-10 years.
- paulpan 5y agoAgreed, though to add, there's also other ARM chip companies that will undoubtedly challenge Qualcomm: Mediatek, Samsung, Altra, etc. Even if the incumbent x86 duopoly of Intel/AMD loses, Qualcomm's track record doesn't inspire confidence. One notable example is their former Kryo custom CPU cores. For a time ~2016, it seems like it gave Qualcomm a distinct advantage to own their destiny and move away from reliance on ARM for their CPU core designs (e.g. SD888 is all X1, A78, A55 cores). But it ended up flopping for various reasons and just like their competitors, Qualcomm was back to using ARM designs. This Nuvia acquisition is akin to having another go - but will it be better than ARM's X* lineup? At the same time, there's 2 additional trends that seem to further erode Qualcomm's attempt: 1. Major Android manufacturers are increasingly moving away from Qualcomm's chips, e.g. Google's Tensor, Samsung's Exynos, Huawei's HiSilicon. This means less revenue for Qualcomm and conversely less money they can throw at R&D. 2. Extreme correlation with chip fabrication technology, aka TSMC. At this point, performance is heavily tied into whether you can secure production on TSMC's latest nodes. No one can outbid Apple.