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 Alder Lake vs. Apple M1
- Thread starter leman
- Start date
- Sort by reaction score
Whether it has an x86 version, ARM version, doesn't matter. As long as it's available on macOS it will run on ARM Macs, better than it did on Intel Macs. And if it does not run on macOS, well, then it's not like you have much choice.
Oh come on, you know very well that it would make a horrible machine. You could do stuff like that years ago when things were dead slow, but it does not fit within the modern hardware paradigm. ARM Macs run Intel software without any issues anyway.
What do you mean?
Same as always. Niche hardware is unlikely to have a macOS version and then even less likely to have a Apple silicon version. If macOS is supported and is being updated, it is likely that there will be an ASi version.
The biggest problem is that third party kernel extensions are being deprecated so support is more difficult. For example there is a kext for my USB-C dock to support the Apple USB Super Drive’s higher power requirement but I don’t use it because I have to reduce my OS security to install it. As Apple adds user space support for various kernel functions this problem will be reduced but right now support is spotty.
No, I know it would make a machine that fit my needs, unlike an M only based machine. As it is, I've already bought my last Mac, and that would be a shame, I want to run a Mac at home.
Like I said, it still exists because it's more versatile, and it's not for performance, but compatibility.
And no, ARM Macs don't run x86 Windows well at all, even in a VM.
The cause of it, no, it works just fine on my Intel Mac, great in fact, and it's my most versatile machine. Apple's change to M1 makes it not work on new machines. (And my M1 MBA is gathering dust.)
Microsoft could make it moot by making a better Windows on Arm and allow me to license it, but they are under no obligation to do so. I'm certainly angry about this, but really, like I said, what do they have to gain by fixing it.
But most of all, it's my problem, and if I want want to run what I need to run, I have no choice, Apple is out, even though I would prefer not. The only reason I'm around here and speak up occasionally is because I still like Macs, and still own a few.
While I agree the current ambiguity of the Arm Windows license is a problem, Microsoft seems not to care. Given the report of Qualcomm exclusivity it looks like Microsoft is applying benign neglect and turning a blind eye to license violations.
And Apple is under no obligation to continue using processors from Intel or AMD. Apple is also not responsible for making sure abandonware runs. Spending way too many resources to cover for lazy developers is one reason why Windows is such a mess.
Well, IMHO, there is no computer that runs x86 Windows well, because .... Windows... ;-)
It would be expensive, slow, inefficient and very buggy. There is no way to make this kind of Frankenstein's monster without significant drawbacks that would render the entire enterprise moot.
That's not enough for someone that needs to stay compliant. I'm an IT Manager and if Microsoft audited us, it would be my a$$ that got fired because of the penalty. You can bet if the licensing guys are talking to you, they wont turn a blind eye, but for normal, non business people, I suspect you're right, at least until they want to start charging for it.
Never said they were, but that doesn't mean I can't complain about it, at least while I still have Mac's.
Now that's funny. Think of money, that's the primary mover, not your hypothetical laziness.
I guess you haven't run Windows in a long time.
Have you ever owned or used a machine like that? My guess is not.
Sorry then.
But I don't share your opinion about Windows, it's actually pretty good these days.
I never ate a **** sandwich but I know it wouldn’t taste good. (BTW, I’ve used Macs with PC Cards in them, and that sucked).
The compromises in such a system would be awful. How would the UMA work? The x86 has full reign on the bus, too? How is memory ordering between the x86 and the Arm enforced? How is memory managed in the first place? How is it allocated without screwing up each processor’s page tables? Just split 50/50 so each gets half the memory, I guess?
It can’t be done in any way that wouldn’t result in awful performance, awful power dissipation, awful security problems, etc.
No, I have not, because they do not exist. I know of some failed experiments attempted a while ago, when it was semi-viable. I hope I understand enough of technology to know that such a hybrid machine has no perspective in modern computing. At most what one can do is a subordinate isolated semi-PC that communicates with the main system via some sort of remote VNC-like interface.
But the constant complaining won't accomplish anything. Do you honestly think that it will cause Apple to reverse course? As the old saying goes, it's like drinking poison and expecting the other person to die.
Yes, it is cheaper in the short term to write a ****** app, toss it over the fence, and then forget about it. I'm sorry that you chose software vendors who do that. Not all do. Meanwhile MS has to bend over backwards to bring all the old cruft into each version of Windows so the bad actors continue to act badly. You may appreciate this, and many do, but there are negative consequences of this as well. OTOH, being proactive and spending resources to update and modernize over time tends to work much better.
Also, if money is your primary mover, you're doing it wrong.
I've used modern adaptions, they work pretty well.
As for the Mac with a PC card, that was back in 68000 days? Not applicable.
That's not how they work, but whatever.
It's *NOT* 2 types of processors in control of the main OS, there's only one system control CPU, and an app and drivers that pass things from and to the sub processor. It's no different than having just another dedicated device that is controlled by the OS. Think of it as a virtual machine that just happens to be backed by specialized hardware.
In the Power9 machine I described, that's exactly how it works, but it goes a little further in having it's own hardware that it talks to too.
They actually still exist, just not on PC level hardware. Doesn't mean it couldn't though.
It doesn't stop others complaining around here, does it? I'm still invested in the architecture, so I have a reason to care. You complaining about me complaining has absolutely no way to change anything either, so why are you?
LOL! Why do you think company's are in business? And since I have to eat, I have to make money too. I'm not into computers for pure altruistic ideals, it's just what I can do to live. Even the open source guys need a day job.
The supposed sub processor still needs to access memory, needs to map its address space, and because you have one OS passing things back and forth (you want to be able to copy and paste? To have apps show up in activity monitor? To access the file system?), you can’t treat it as a stand-alone processor that doesn’t need to access and coordinate the same memory space.
What “modern adaption [sic]” have you used that supposedly works this way?
This has been my experience too.
Even compiling C# .NET 6 on JetBrains Rider EAP on M1 Max, it was performing faster than our typical Ryzen 5950x workstations. For me and my workflow, that’s game changing. That’s a laptop part with 8 performance cores, running at a significantly lower clock speed and higher IPC, beating out a desktop workstation part with 16 performance cores in a fraction of the TDP. And I can carry it with me on a plane, move it easily to the kitchen while I drink my morning coffee!
On a separate note, I have been exploring more of the Accelerate framework and retooling parts of an existing code base to utilize accelerate as I port what was originally a Windows developer Visual Studio app with SSE optimizations over to XCode and M1. So far performance improvements on both i9 Intel MBP and also M1 have been significant when I started to tap into Acceleate - but I’m seeing even greater performance bumps on the M1 relatively so I suspect it’s a combination of additional accelerators and co-processors there getting utilized. The “downside” is that I’ve really ended up now with two separate and succinct projects to maintain - one for Windows and another for Mac OS!
May I ask , what numerical workload code are you using specifically on x86 that uses AVX2 and AVX512? It must be very specialized? Have you investigated ported to see if you could achieve a similar level of performance on M1/pro/Mac equivalent via Accelerate, Metal or other proprietary library from Apple? I know from your other comments around here that you’re somewhat of a code junkie, so no doubt you’ve explored lots of options and have a really good reasons.
I’m just curious if your particular usecase and workload just lends itself better towards x86 and whether this is a hardware limitation from your perspective, a software issue - i.e. you’re using an open source BLAS library or something that is not well optimized yet for AS or a combination of both
Just interested to hear your take
Hope you are staying well and keeping safe.
Happy Holidays
Last edited:
You allocate your memory on bootup of the App, and it keeps that address space while it's running. But you could actually put the RAM on the board with the x86 processor.
Of course, the app handles that, just like RDP and virtual machine software does.
No on that, why would I want that? Like I said, it's not running as a Mac OS, and Windows has it's own task manager. If you need to shut the whole x86 processor down, one task in the Mac activity monitor is enough.
That's because it doesn't have to! Again, think of how virtual machines work.
IBM Power9 (and Power10 now) IBM i midrange, with attached x86 server. The x86 server has it's own RAM, but the host machine serves the disk, devices, control, backup.