Macs With Intel Processors Still Coming Amid Transition to Apple Silicon

[Boil]

The MCM method gets expensive quickly. Discrete graphics isn’t power efficient, but it is powerful. And some professionals simply need more than 256GB RAM, regardless of what myth some people may believe with regards to memory on Apple Silicon.

LPDDR5X will bring 64GB chips, so a theoretical quad SoC ASi Mac Pro could have up to 1TB of RAM, with 2TB/s UMA bandwidth...?

Would that be the reason why we didn't see any high end minis? Maybe they will replace those with they mid-tier Mac Pro, with all the M1 Max dies.

So replace a machine that costs about US$3k loaded (not counting excessive SSD pricing) with one that starts at US$6k with much worse specs...?

Maybe they can finally create that mythical xMac and have it be the very final Intel Mac Pro! I can imagine... x for Max, then Mac for, well, Macintosh.

iMac (24")
iMac Pro (27")
xMac (max, buildable)

The final Intel Mac Pro will be new internals, but the 2019 Mac Pro Cheesegrater 2.0 will remain the chassis of choice...

"No xMac for you...!" (Seinfield / Soup Nazi)

I see this rumor being true for a few reasons.

1. The "new" Mac Pro design will coincide with its transition to Apple Silicon. It's the only conceivable approach - why do a re-design for the Intel chips but not your new "flagship" machine? Additionally, it's clear the M1-series of chips require significantly less active cooling than the current generation Intel Xeon chips. A redesign for a new cooling mechanism makes the most sense in this regard.

2. The Intel Mac Pro update will be a silent/low-key refresh: new generation processors, some minor spec adjustments, and that's it. Would make sense to have a last alternative for customers wanting an Intel device (for software/hardware compatibility, etc.)

I could see a press release for new internals on the 2019 Mac Pro, issued before WWDC 2022, then the new smaller (half the volume) ASi Mac Pro is debuted / showcased at WWDC 2022, shipping by the end of the year...

 

[Le0M]

I always thought Apple designed M SoCs so as to combine many of the same chip together to form a much more powerful one.
If the Mac Pro will only be twice as powerful as a M1 Max, then it won't be as great.
It has to be 10+ M1 Max, for one Super Sayan SoC.

 

[JPack]

One can only presume that when Apple release the M1 QuadMax/DuoMax, it will not only have up to 256GB RAM on SOC, but will also have the ability to add extendable external RAM up to, at the very lease, 1.5TB, but I'd guess Apple would want to blow that away and allow much more than that. I mean, 32 bit had the limitation of only being able to address 4GB, but 64 bit can address up to 18 Exa Bytes, i.e. 18,000,000 TB, so that's the upper limit at the moment, before we'll have to move to 128 bit architecture.

DIMM slots would starve the GPU of memory bandwidth. You would need 8 DIMMs just to get 400 GB/sec.

There's a big cost to putting the GPU on die - you need to rely on package memory.

 

[sideshowuniqueuser]

DIMM slots would starve the GPU of memory bandwidth. You would need 8 DIMMs just to get 400 GB/sec.

There's a big cost to putting the GPU on die - you need to rely on package memory.

So you're trying to suggest they will simply stick to 256GB max RAM on the AS Mac Pro? Somehow I doubt it, they'd be laughed out of the room.

 

[SFjohn]

name me one company that goes "oh no. that employee has been using adobe premiere for 8 hours today on the $15,000 mac pro. i am worried about the electricity bill for this month is going to cost a few dollars more! noooooo0o0o0o00o0o!!!!!!!!!!"

Sever farms spend an enormous amount of $$$ to run & cool themselves. If they can increase performance & lower cooling & energy costs they will get on board.

 

[JPack]

LPDDR5X will bring 64GB chips, so a theoretical quad SoC ASi Mac Pro could have up to 1TB of RAM, with 2TB/s UMA bandwidth...?

Maybe for 3nm and below. With 4nm, such a dual-die MCM package would already use 200W.

 

[Malus120]

I think this makes a lot of sense and I just bought a new M1 Max MBP.

The reality is there are still a good number of people/businesses out there who need a macOS system that
1. runs on x86
2. is highly customizable / upgradable
(at least for the time being)

The M1 was a great window into what Apple Silicon could do, and the M1 Pro/Max have demonstrated Apple can scale up the design to meet the needs of the Pro laptop space but I'm not sure Apple Silicon is ready for the Mac Pro.
1.The memory, while fast, is non standard, not expandable, and currently tied to the SoC configuration.
2. The supported display configurations are somewhat limited and tied to the SoC GPU.
3. PCIE lanes seem to be somewhat limited, non Apple (e)GPUs are currently unsupported (and may never be supported,) and support for other PCIE devices via Thunderbolt can be hit or miss.
4. Macs will always have to compete with the iPhone for chips, especially at the bleeding edge
Not saying these and other issues couldn't be resolved (most definitely could) but I'm not sure Apple is ready or able to at this time.

The current Mac Pro is the result of Apple finally listening to Pro users about what they actually want in a professional oriented desktop Mac, after more than half a decade of just letting the 2013 nMP, which wasn't what a lot of pros wanted in the first place, languish in purgatory. If Apple were to build an Apple Silicon Mac Pro today I have a feeling it would look a whole lot more like the 2013 nMP than the 2019 (n)nMP, and that's just not what people want IMHO.

I think there's been a logical progression here that Apple should stick to
4Q2020 - Consumer facing MBA/13" "low-end" MBP, Mac Mini
1H2021 - Consumer facing 21" iMac replacement (24" iMac)
4Q2021 - Pro Laptops

1H2022 - Prosumer iMac/iMac Pro and Mac Mini (?) + Second Generation consumer laptops/desktops
4Q2022 - Second generation Pro Laptops
2023-24: Mac Pro?

Waiting until at least 2023 for a Mac Pro makes sense. The software support will be even better than it is today and they will have had plenty of time to design a new chassis and decide how much they want (or the market wants them) to support open hardware standards. They can also see how things play out in the x86 space. Maybe AMD or Intel is able to take things in a direction in the high power space that Apple Silicon (which is designed first and foremost to be low power) just cant go, in which case they could maintain the split indefinitely.
For those who are sad because they really want an Apple Silicon pro desktop... don't be, I'm pretty sure we'll still be getting one or more in 2022, and they'll likely be as good as any Apple Silicon MP Apple could've released in 2022 anyway.

 

[TVGenius]

No way in hell I'm buying any new Intel hardware unless up front they'll give a minimum window for Rosetta etc support to continue. Why should buyers feel comfortably investing in hardware that's more likely to be used for important functions when it basically has a ticking countdown on it until it won't receive OS/security updates or might not work with necessary apps?

 

[Boil]

One can only presume that when Apple release the M1 QuadMax/DuoMax, it will not only have up to 256GB RAM on SOC, but will also have the ability to add extendable external RAM up to, at the very lease, 1.5TB, but I'd guess Apple would want to blow that away and allow much more than that. I mean, 32 bit had the limitation of only being able to address 4GB, but 64 bit can address up to 18 Exa Bytes, i.e. 18,000,000 TB, so that's the upper limit at the moment, before we'll have to move to 128 bit architecture.

DIMM slots would starve the GPU of memory bandwidth. You would need 8 DIMMs just to get 400 GB/sec.

There's a big cost to putting the GPU on die - you need to rely on package memory.

The "secondary RAM" could be seen as a faster swap than going to the SSD(s), it would not hamper GPU performance with the "primary RAM / UMA"...

So you're trying to suggest they will simply stick to 256GB max RAM on the AS Mac Pro? Somehow I doubt it, they'd be laughed out of the room.

LPDDR5X allows 64GB chips, so a quad Mn Max configuration could have up to 1TB of RAM with up to 2TB/s of UMA bandwidth...

But LPDDR5X might be for M2 Max SoCs; Apple may very well release only a M1 Max Duo in a new Mac Pro Cube chassis for the initial Apple silicon Mac Pro...?

Maybe for 3nm and below. With 4nm, such a dual-die MCM package would already use 200W.

So...? Power limitations do not matter for a desktop workstation...

No way in hell I'm buying any new Intel hardware unless up front they'll give a minimum window for Rosetta etc support to continue. Why should buyers feel comfortably investing in hardware that's more likely to be used for important functions when it basically has a ticking countdown on it until it won't receive OS/security updates or might not work with necessary apps?

If you are running an Intel Mac, Rosetta is not a consideration...?

 

[Shirasaki]

then what is your point?

I think I missed the point too.would u mind explaining us what you meant ?

For a small scale, one or two super powerful computer doesn’t seem to cost much more of energy, but the total can quickly add up if you neglect it. Besides, for bigger businesses that can afford a fleet of those powerful computers cannot just set aside a certain amount to pay power bill without thinking twice, as governments, environmental activists, media and maybe also general public could create noise on why the energy consumption is that high, forcing companies to allocate PR resources to address them.

Needless to say, it is much better to hop on this “protecting environment” train than not boarding this train, both in the short term or in the long term.

 

[ClevelandGuy]

I don’t know if this has been said yet, but Apple is still using an Intel Thunderbolt controller chip in Apple Silicon Mac’s. Unless/until Apple moves this into the M chips they’ll still be an Intel customer.

They are not using Intel’s Thunderbolt controller chip in the M series Apple Silicon, the Thunderbolt controller is part of the M chips themselves.

 

[Shirasaki]

If you have thousands of workstations, you are talking about the concerns of a large business, not a "professional". I don't think anyone would disagree that a a business with thousands of workstations wouldn't care about their very sizable electric bill. But for a business that size, IT management can exploit economies of scale that just aren't worth it for a single professional or small outfit.

Economy of scale or not, what’s limited is still limited, as there’s only a set amount of capacity the business could use at any given time. For small businesses that has one or two of those workstations maybe not a huge concern, but still a concern nonetheless.

 

[Shirasaki]

Mac with M1 has worthless library of software. This is even worse for actual pros and not Youtube pretenders.

And Intel's latest Golden Cove CPU is superior to everything in the market right now. If they can get Xeon based on that architecture, then there's really no reason to glue more slow cores on the M1.

Well well. I mean, M1 software library is still lacklustre, and will continue to be so for the next year or two, but to call M1 slow? Besides Intel themselves, I feel anyone wiling to call M1 slow would really need some courage, more so when Apple “courageously” removes the headphone jack (and sell AirPods conveniently at the same time).

 

[cmaier]

I don’t know if this has been said yet, but Apple is still using an Intel Thunderbolt controller chip in Apple Silicon Mac’s. Unless/until Apple moves this into the M chips they’ll still be an Intel customer.

Hasn’t been said because it’s not true.

 

[maxoakland]

No surprise here.

The MCM method gets expensive quickly. Discrete graphics isn’t power efficient, but it is powerful. And some professionals simply need more than 256GB RAM, regardless of what myth some people may believe with regards to memory on Apple Silicon.

Also expandability is important. That’s one of the best things about the Mac Pro. And I miss it in their portable and consumer systems too. A lot

 

[loby]

I get the feeling that if Apple intended to release any additional new Intel-based Macs, they’d have done so before the end of this year. Perhaps they intended on doing one more Intel-based Mac Pro refresh in January 2021 when the report from which this article stems was published but opted to cancel it.

Releasing an Intel-based Mac Pro refresh too close to the Apple silicon-based redesign, especially if the former remains available for purchase after the latter is announced, might read as a lack of confidence in the latter.

“but it seems that Apple is not yet entirely confident enough to transition the ‌Mac Pro‌”

It is probably not based on confidence, but business and contracts. If anyone remembers when Mac Pro 2019 came out Tim Cook talked about it being upgradable etc. can’t just ditch it after all the hipe (though Apple has done that in the past). Probably still has contracts with Intel that need to run out first before going total silicone.

Besides…it makes sense to have both Intel and silicone on Mac Pro for those who bought into the 2019 version.

Many need windows/bootcamp because reality workflow with many audio/video production firms is tha they use both Windows and Mac applications, especially Hollyweird.

 

[DeepIn2U]

Goooood….Makes my Hackintosh Viable for years to come 🙃

Oh no.

Honestly I wish Apple invested the resources to better improving Rosetta 2 or even better grant eGPU support for the latest cards as well as virtualization or emulation support natively, but what do I know.

MBA, MBP, iMac and Mac mini have ALL transitioned to Apple Silicon. Honestly I don't think Apple's decision here has anything to do with performance of Intel. I think it's more about fulfilling a contract agreement to not renege out and have to pay needless contractual breach to pay Intel.

Does anything think pro users really want another intel-inside Mac Pro, or would they prefer all plug-ins etc for Audio pro apps and other apps in various pro fields to be fully written and supporting Apple Silicon ??! Especially with an Intel Mac Pro which will be eclipsed within 2-3yrs after you've spent some $10K+ not a good investment for long term use.

PS: I type this on a used Intel MBP 13" I bought 2mths ago used because it was the best bang for buck that'll be supported for the next 2yrs, which will be gifted to my sister after I upgrade to an M1 Pro based MBP. Either way I don't see Apple making poor choices/decisions here.

 

[segfaultdotorg]

Since Apple never formally adopted the "Intel Inside" marketing badges, I propose they laser-etch the "Intel Inside" logo so that it is visible on every side of the product.

 

[Neodym]

Let’s hope they don’t design themselves into a corner again with the new, smaller case … ;-)

 

[dysamoria]

Makes sense, really. Anyone buying that kind of configuration probably has very specific and critical software in mind. If Apple releases Mac Pros for both architectures next year, that would buy time for such software companies to transition from x86 to AS without making their customers bug test software in a production environment. Even just being the last Intel Mac, we don’t know how long it might stay in production. The Trash Can Mac Pro ran for many years with no significant design changes, and this last Intel Mac might do the same in order to keep a niche customer base with deep pockets.

Apple would have to continue updating an Intel Mac Pro for years to come, or the studios would abandon them for PCs again.

 

[dysamoria]

You’re joking right? The new MBP Pro and Max are blowing a lot of high end current MP machines away for working pros. Performance/Watt is where it’s at.

You’re not running a VFX company, are you?

 

[svish]

Might see the last Mac with Intel being launched in the first half of 2022.

 

[Planetary]

This plan is stupid. Nobody is going to buy these.

 

[dysamoria]

Or there are customers that still need Bootcamp with Windows capability, maybe that's why we'll see extended refreshing on Intel CPUs. I use Bootcamp for gaming and have no intention moving to Apple Silicon without Windows capability.

I’ve been waiting for Apple to release a proper Mac Pro and retina display for a decade. I want to have a Mac Pro to replace my iMac and 2008 PC. Bootcamp for Windows gaming, Mac OS for everything else.

Then they kept not releasing a reasonable machine, and no Apple retina display. And then they finally released a proper Mac Pro, but priced it entirely out of reach of even enthusiasts. The base $6K Mac Plutocrat is poor value, and its $5K display is overkill. Then came the end of Bootcamp via Apple Silicon.

I don’t see them changing course on the Intel Mac Pro pricing, but I’m open to them doing so.

 

[SFjohn]

I think it will be fun & interesting to see how the the new Apple Silicon Mac Pro compares to the new Apple Intel Mac Pro in 2022.