And exactly how many of the widely used consumer applications need that? Don't forget that Apple didn't just yank all the Intel machines off the shelve and replaced them with ARM. There's a plethora of Apple machines out there still being sold with Intel CPUs. Get those until there is a 100% coverage. There are even 8 and 10 year old Macs out there still in use, it's not like they expire like milk. You're more than covered transition-wise. Not sure how many other computer manufacturing companies give you that level of transition flexibility.
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.
Intel Urged to Take 'Immediate Action' Amid Threats From Apple Silicon and AMD
- Thread starter MacRumors
- Start date
- Sort by reaction score
And exactly how many of the widely used consumer applications need that? Don't forget that Apple didn't just yank all the Intel machines off the shelve and replaced them with ARM. There's a plethora of Apple machines out there still being sold with Intel CPUs. Get those until there is a 100% coverage. There are even 8 and 10 year old Macs out there still in use, it's not like they expire like milk. You're more than covered transition-wise. Not sure how many other computer manufacturing companies give you that level of transition flexibility.
I understand, but that wasn't the original point that was being made or responded to.
I find it interesting that people are already signing Intel's death certificate. I guess they assume that Intel has no alternative to Apple Silicon. Maybe, just maybe, they have one and have been riding the x86 gravy train until it pulls into the station.
Intel tried to move away from x86 multiple times (cf. i960 and Itanium). Their strength has ultimately proven to be in compatibility, not in modern ISAs.
That’s not how things work. Intel doesn’t secretly have 7nm fabs up and running smoothly at high yields, just waiting until they had a reason to use them. They don’t secretly have a RISC architecture without the need for 4 decode pipeline stages with microcode roms and a state machine, that they were just waiting for the right time to spring on the world. And it takes 2 years from when the architects sit down to define the chip until the chip is taped out, so if they started when M1 was announced, it will be awhile.
And even if they did have a RISC architecture ready-to-go (despite having sold off StrongARM, which would have been pretty useful to them right about now), it will never compete with Apple Silicon, because Apple Silicon is targeted to a single manufacturer’s machines, where that manufacturer controls the compiler, the OS, and all the other hardware. Intel needs to make the lowest common denominator for everyone else.
exactly, and so does AMD, for the current PC (laptop, desktop) market.
and there is no-one in the Arm market that could rise as the next "Intel", QCOMs Arm processor in the Surface (Windows Arm) have, meh, performance at best, not enough to be serious competitor to x86. There is no-one in the x86 world that will come out with a similar integration like apple anytime soon, and when/if they do, who is going to produce all those chips? fab capacity at the low end nodes is extremely tight ... Intel and AMD will continue to drive x86 and are far from "doomed".
Apple has better management than Intel. Apple still has room for improvement.
Precisely. And from the article I posted Intel suggested people stop overclocking a 7th gen i7 on an overclockable CPU to help with the heat. I guess Apple could just NOT put in an i9 in their iMac, but people would complain that Apple doesn't provide more cores blah blah blah. Its a no win situation.
The ONLY thing Apple should have done is brought the iMac Pro cooling down to the iMac. Which would help a little. But like you said the i9 is 296 watts.
I think you are blowing out of proportion the "rush to port". Since Apple had announced the shift to Mac Silicon, much of the porting efforts were probably already on developers' roadmaps. I suspect the ones that you think ported quickly were devs/teams that either have a rather large Mac user base or had very low friction points in creating the universal binaries.
See above. Your concept of investment is inflated.
I'm not sure what your point is regarding ARM linux distros. Apps on Linux ARM platforms are typically very spotty, let alone a significant learning curve for anyone that isn't linux/unix savvy. Microsoft has been working with ARM for their Azure servers and Surface PCs for a bit now, so this likely only accelerated further spending into that segment.
The very significant detail that you are forgetting is that Apple owns the SoC/hardware and the OS/software. Is there another vendor out there that owns literally the full stack? Microsoft in their efforts have been collaborating with QC.
Don't get me wrong, I'm very excited to see this type of competition as it pushes technology advances, but I'm also tempering my expectations.
Just to give you some additional points to mull over:
- x86 has been around since 1978 (42 years so far). The average life expectancy of a U.S. man is about 76 years.
- Windows 95 on 32bit x86 for everyday people released 1995 (deprecated 2001)
- Windows Vista on 64bit x86 for everyday people released 2007 (deprecated 2017)
- Windows 10 for 32/64 x86 and ARM for everyday people released 2015 (5 years so far)
- Apple on PPC first released 1997-2011 (14 years)
- Apple on Intel first released 2006 (14 years so far)
- Apple on ARM first released 2020
You can be fantastic and have room for improvement. You can always improve, but Apple's among the best or the best.
I wasnt sure we can use the "Stupid" word isnt it rude and against the regulations on Macrumors?Merced was x86 compatible. It failed because it was a stupid architecture.
INTEL REVEALS DETAILS ON HOW MERCED WILL BE X86-COMPATIBLE
Santa Clara, CA -- As Electronic Engineering Times reported, a recently issued patent and new information from Intel Corp. provide fresh insight into a technical mystery surrounding the company's upcoming […]www.hpcwire.com
@sracer People at signing Intel’s death certificate not because of today, it’s because of what today means for the future. For big future facing tech companies like Boeing, Pfizer, and Intel, the money they are making today are based on the successes in the years prior. TODAY is Intel’s future “years prior”... and Intel aren’t showing anything currently that will lead to the same leadership position they have today. It could be that, two years from now, Intel has a breakthrough, BUT as cmaier indicates, THEN there’s more years before the breakthrough actually reaches the market.
The death certificate is because even if Intel was doing things PERFECTLY right now, they’d still have a hard time ahead of them. They’re not executing perfectly right now, so their outlook is grim.
To paraphrase Stephen Colbert: Intel's glass is 2/3rds empty. That last third is usually backwash.
I don't think people are evaluating the market based on just the M1. Third Point didn't-- they have a laundry list of concerns.
I think a lot of people see the eventual demise of Intel as inevitable and probably overdue. We've known that x86 is a challenged architecture, that Intel's management has been a mess for decades, that their principle advantage was their process, etc. We've watched them fail at every attempt they've made to diversify their product line, lose ground to AMD on design and to TSMC on process.
If this was still just Intel vs. AMD+TSMC then I'd agree it was just a matter of catching up. In the end they're bound by the same constraints so they should be able to converge to equally effective designs.
But it's not just those players anymore. Apple Silicon not only added instruction set architecture as a free variable to optimize, but also system level architecture. The solution space just got a lot bigger and the x86 subspace just looks undesirable in this new world.
The PC architecture is based on modular commodity design, where commodity parts are interconnected by defined interfaces. This allows competition and innovation in each subsystem, and allows system vendors to mix and match among vendors. The exception, of course, is the processing subsystem which was always x86. This was great for innovation in system design over the past few decades, and also great for keeping Intel in the driver seat since they always had a scale advantage over AMD.
That doesn't seem to be the best system tradeoff anymore. It makes it impossible to optimize across the defined interfaces.
Furthermore, the x86 architecture really isn't what anyone would choose if they were starting fresh. It's too ugly to maintain. Arm may not be perfect, but it's certainly better and it's sufficiently mature. The performance difference has gotten extreme enough that even the backwards compatibility argument starts to fail-- backwards compatibility is maintained through translation with sufficient performance.
And backwards compatibility just isn't the challenge that it used to be. Once Windows, Visual Studio, and its APIs port to a new architecture (not necessarily Apple Silicon, but the inevitable PC replacement architecture), the vast, vast majority of applications will follow fairly easily. Almost everything is done in higher level languages these days.
If Intel had shown any ability to adapt in the market in the past, or to design anything other than x86, then maybe people would give them better odds, but Intel's history ain't great in these areas and the fact that they hadn't anticipated this move in the market just further supports that assessment.
Could Intel cobble together a new architecture in response to M1? Sure, but it would be brand spanking new without any of the design maturity and market adoption that Arm, and specifically Apple Silicon, have. @cmaier estimated 2 years from design to tape out. That's 2 years for a 1st generation part to reach the market and compete with what will likely be 7th or 8th generation Apple Silicon designs?
I don't know that there's anyone out there right now that can compete with AS directly in the foreseeable future, but they don't need to. They need to put out a PC alternative to AS, and AS may be the benchmark that they're compared against but perhaps will remain a different market segment. Regardless, Intel is at a standing start as far as I can tell, and given their history I don't see any reason to expect they'll win that race.
In short, I don't think what you're hearing is entirely a case of people reacting to a single event-- I think people have a much more sophisticated view based on a fuller understanding of the technologies, players and economics involved. Like I said, I wouldn't count Intel out entirely, but this is a totally different situation than "OMG Threadripper!!! Intel is dooomed!!111".
In the intermediate term Intel doesn’t necessarily need to compete with Apple Silicon because overwhelmingly, desktop and server software runs on non-Mac x86 operating systems and processors.
(Also Macs and iOS have been good for a long time and plenty of people don’t buy Apple products, and many will continue to not buy them for other reasons even beyond the software landscape - such as their perceptions of value and openness and even politics)
Intel does need to compete with AMD and AMD and Intel pricing is converging just as market forces would dictate. Intel still has higher gross margins than AMD, presumably those will continue to shrink indefinitely or at some point Intel will become more competitive. Clearly as of not many years ago Intel had better real world processors than AMD. So it seems at least plausible that Intel would be capable of solving problems since they have in the past.
Also while it appears Apple is in a great situation now (which is not a secret by any means) it doesn’t take much imagination to see many of those advantages become less important in the not too distant future (which is not a secret by any means)
As to non-Apple ARM servers and PCs, between what Linus and Microsoft seem to be indicating, as well as inferring from the pace and nature of change in servers and PCs in the past, it seems like there is plenty of time remaining for x86.
Last edited:
This is the kind of thinking that will sink a company... Yes, they do. Part of what got Intel into this situation to begin with was thinking they didn't need to compete with Arm, because "Arm was for embedded applications".
Depending on the benchmark you're looking at, the M1 is too far ahead to ignore. For example, this benchmark suggests the M1 MBP is about 60% faster than the Intel MBP running Handbrake. Intel seems to show what, 10% improvement per year? That puts the M1 5 years ahead of Intel.
Moving from Intel to M1 will instantly put a user 5 years ahead of the their competition technologically. Not many businesses are going to put up with that competitive disadvantage for long.
Last edited:
Isn't it too late? Steve Jobs offered Intel to start producing custom ARM chips more than a decade ago, but Intel refused. That's why Apple was forced to start designing processors in-house.
Now, custom Apple M1 chip has reached the point when it outperforms the best Intel chips. What we have now is the top-line Intel chip consumes 125W and comparable performance Apple M1 chip consumes just 5W. That's x25 power consumption difference! The main growth driver in Microsoft, Amazon and Apple are cloud services. The biggest bill for any cloud data center is power consumption and cooling. That's why those companies quickly move to much less power-hungry ARM processors!
Now, custom Apple M1 chip has reached the point when it outperforms the best Intel chips. What we have now is the top-line Intel chip consumes 125W and comparable performance Apple M1 chip consumes just 5W. That's x25 power consumption difference! The main growth driver in Microsoft, Amazon and Apple are cloud services. The biggest bill for any cloud data center is power consumption and cooling. That's why those companies quickly move to much less power-hungry ARM processors!
One should not forget the lesson of the iPhone. Thanks to the iPhone, Apple has aggregated the best customers. And now with over a billion active iPhone users, this is a large customer pool that Apple can market additional accessories, apps and services to. This in turn also means that android smartphone OEMs are stuck in a race to the bottom, which means lower profits and consequently, less money available for R&D, while the apple ecosystem grows stickier and more established with each passing day.
The end result is that while the iPhone loses in market share, they end up dominating in just about every other metric which matters.
I can see the same situation happening to the PC market. Up till now, Intel would have been indifferent as to whether consumers bought a PC or a Mac, so long as they all ran intel chips inside. But now, every additional M1 Mac sold is a computer that Intel doesn't get any money from.
Will we also start seeing a similar situation in the PC market where the more well-heeled customers start buying Macs because of the huge disparity in performance, and because supposed drawbacks like the lack of upgradeability simply doesn't matter to professionals like doctors and lawyers? Leaving only the $500-700 dollar price range for PC OEMs to fight over. Lower profits on laptops means less money for innovation.
Classic Apple doing what Apple does best: take an emerging product category with a frustrating user experience and deliver a polished product made possible by its control over both the hardware and software.
I'm with you - remember on the Retina MBPs when Jobs removed the DVD drives, people freaked, and he said the way of the future is app stores, and disclose software installs? People thought he was nuts, he was mocked. But look how that single move drove the industry forward.
Not sure what you’re “remembering”, but when the rMBPs launched, Jobs was dead.
Absolutely true. And parenthetically I saw somewhere the market share of Macs are gradually rising, so it's a double whammy.
The race to the bottom is somewhat true even before the m1 chip. However, I don't know what would drive the masses to make a purchase decision between m1 and wintel. Speed and/or performance may or may not be a factor. 100% windows compatibility may or may not be a factor. Those more well-heeled customers who already have a mac and want to upgrade may do so.
Remember Y2K. Multi-year effort costing billions and billions. There is no such thing as quick, and clearly not cheap. And all this is mitigated is 100% windows compatability is needed.
Hindsight is great, but that is exactly how the market panned out. Apples' m1 chip would never have made it the way Intel x86 architecture did. Because software at scale has to be support for 30 to 40 years. Something Apple does not do and history will show the Arm in 40 years will be the next x86.
And the hardware/software market can be fragmented if each company that has time, resources and budget develops their own chip and operating system. History has a way of repeating itself.
Doubtful. Much of the medical software in the US is for PCs. There are also a large amount of deals made with hospitals to supply these PCs. It’s going to be a bit until Mac Silicon is officially supported in an enterprise / corporate environment. Lawyers aren’t gobbling Apple laptops.
If you are in the Apple ecosystem already, that is where this excels. For your teaching and semi prosumer needs, it will more than satisfy. I don’t see a large portion of existing PC users jumping to Macs due to these benchmarks. That is a different use case to solve.