Apple just announced all their computers will be ARM based by the end of 2021. This will effectively kill BootCamp as a viable option for running Revit, and Virtual Machine programs like VMWare and Parallels will have to emulate hardware causing significant performance hits. So, Autodesk, you're about to lose a significant install base. Will this finally get you to make Revit a native Mac app after 14 flipping years?
@kylem1701 wrote:. So, Autodesk, you're about to lose a significant install base. Will this finally get you to make Revit a native Mac app after 14 flipping years?
no why would they
do you have facts supporting losing a significant install base?
ARM processors aren't very powerful and it is still doubtful they can actually create processors replacing the Intel server processors they use. Look at Anandtech for a recent article. They have a Rosetta 2 engine that "translates" the x86 software to ARM. But that may not be suitable for a complex software. At this point it isn't even clear if the MS software will get an ARM version.
I doubt there is a large number of Mac users using Revit. And I rather have Adesk work on bugs and implementing our wish list items, than creating a version optimized for a weak (compared to actual desktop processors) mobile processor. And then what, in 5 years Apple goes back to Intel, or AMD, then they throw away their effort? And Apple takes a 30% cut and forces users to only buy software from their app store. I doubt Adesk likes that part.
Revit is a bit more than a "fart noise" app.
Obviously there are enough Autodesk Mac users to need a Mac version of nearly all their other apps. I work in an SF architecture office that is Mac only. Most places we work with are Mac only. Anecdotal? Maybe. But it’s not hard to believe that creatives prefer Macs and offices in major cities use them. Now those offices will have to decide between Macs in their next upgrade cycles or a single app Autodesk refuses to port because... why? You tell me.
That’s why Adobe, famous maker of fart noise apps, has already ported their entire suite to the processor platform, right? No need to turn this into a petty 90’s style OS flame war. We’re passed that already.
Your response starts off reasonable, but it takes a nose dive pretty fast. It's amazing how upset people get when someone is interested in something they aren't. Mac users don't like windows just like you don't like Mac. No one wants to buy a second computer for one application. Asking for a port is not a desire to "rile" anyone up. I have no idea why you would see it that way. Am I suggesting Windows users are lacking in anything? Have I offended them in some way? And what difference does it make how long I've been in a customer support group? I had a question for the first time and I signed up and asked it. Again, why does that bother you? What about what I'm asking is trolling? SMH.
AFAIK this will be a 2 year transition. Apple even will still introduce new x86 computers over the next 2 years. So it is all speculation on how powerful that will be, and if the high-end Macs will get ARM before the 2 year transition. So it isn't clear how many Macs will actually be in the wild with ARM.
Apple also can change their mind within the 2 years, depending on what trouble they run into or if Intel/AMD come up with a really great x86 alternative. Main reason apple is doing that they are not happy with Intel progress over the last years. This may be a large exercise to put Intel on notice and to keep options open.
IF Adesk makes an ARM version depends on how THEY think the demand will be. No one on this forum will know.
I know most engineering software exists on x86 only. Maybe architects like Macs, but for engineers that is unlikely.
It also will depend on if the add-ins will need to be converted to ARM. That may be the larger hurdle since those smaller companies are important for Revit, but may not have the resources to maintain twice the software.
ARM is more a license scheme and licensees can make their own versions. This gives us many manufacturers unlike the Intel/AMD duoploy, but it isn't standardized.
It is a lot of work to make a software take advantage of all the CPU features. It has to work optimally on all modern AMD and Intel CPUs with a multitude of windows versions. Maybe it is reasonable to make an ARM version, but not an optimized one. Then it may just be easier to use the emulator.
Anyone who can make a good prediction has signed an iron-clad NDA 🙂
Again, I rather have Adesk focus on one version with fewer bugs and better features than to get tangled up in multiple versions that kind of work half-good. Look at the wishlist on this forum and how little of the great ideas has been implemented yet.
For what it's worth, I used to work for a big multi-national who was a big user of Revit. They chose their OS and hardware based on the software they needed to use (Revit, AutoCAD, and a few others), not the other way around, so they had large numbers of Windows boxes. Some of the 'creative types' had Mac laptops but they had little to nothing to do with the engineering/design side.
There are users in the creative industry (read architects) that require Adobe products also, which just imho works better/nicer on a mac system. If Revit is not compatible anymore for them it is likely Autodesk will loose those mac users who will switch to alternatives, Archicad for example, which runs on both operating systems fine (...these two should really work on better interoperability btw). Percentage wise it's probably not that many users affected, but of course I don't have the numbers... Only thing I can think to say on the matter is that it's called Revit AEC collection, not Revit EC, but in the end it's up to Autodesk for which users this is and/or remains true...
Autodesk really has little choice. Shortly all their Mac software will need to be re-written or discontinued. If discontinued most Fusion 360 users would move to Onshape.
But can Autodesk port CAD software to Arm and get reasonable performance. I doubt it unless they can find a way to multi-thread and take advanced of something like 6 or 8 Arm cores.
It would be easier for Autodesk to port existing Mac apps to Linux
As a Mac user I'm thinking about abandoning the Mac. I suspect many others are also going to move away from Macs. In may case Linux works well as I am a long-time Linux user and software developer on Linux/UNIX. I got to using MacOS because it is in fact UNIX.
All of us who would be leaving Mac for Linux are forced to using Onshape. I hope Autodesk to reevaluating their lack of suport for Linux.p
We must remember that Apple is not moving to "Arm" but to "Apple Silicon" which has ARM cores but in addition to ARM coarse, there are a half dozen other cores. It is the other cores that will give the new Macs their speed advantage.
If Autodesk simply recompiles there single-core Fusion 360 app to ARM it will be dead-dog slow. But if Autodesk takes the time to change the code so that it can use something like 8 ARM cores and the other GPU-like cores using Apple's APIs then Fusion 360 can be very fast on the new Macs. But more then a simple re-compile will be required.
So, depending on how much time and money Autodesk invests the result will be either very good or very poor.
Again the way to understand this is that Apple silicon includes ARM cores and a lot more. To run fast on Apple Silicon you need to use multiple ARM cores and those other cores too.
More technically software spends a lot of its time running short pieces of code that are inside nested loops. Apple's plan is to place hardware on the Apple Silicon chips that implement common algorithms in hardware. So the piece of code inside the loop might run 200 times faster. But only if Autodesk re-writes it to take advantage of the hardware. A simple re-compile would not do this.
It is the same with using all those ARM cores. The software can be faster if it can use many cores but a simple re-compile does not do that. Developers have to think hard and work to use those cores.
It seems Adobe took the time to re-write. Will Autodesk invest in Apple Silicon or take the cheap way out? We will have to wait and see.
So here we are, months after this was originally posted, and how the potential landscape has changed! The benchmarks of the M1 are roughly on par with the 10th gen i9 in the 16" MacBook Pro. The performance of "real world" app tests in heavy processor loading workflows definately shows the M1 "holding its own" against more powerful x86 proccessors. I've little doubt the M1 is capable of running Revit. There ought to be AutoCAD for Mac on Apple Silicon performances test *somewhere* out there. So here's the real issue: The Business case for Autodesk making Revit for Mac.
The AutoCAD verticals were never ported to Mac, which indicates the type of user is at play. I don't doubt design professionals and technical professionals are using AMAC, it's the more "hard core" AEC professionals using Revit, and the overlapp on the Venn Diagram with that demographic AND Mac users are likely small. Add to this the number of NEW Revit users Autodesk hopes to gain by rewriting Revit for RMAC, and the business case seems limited to me.
What is of more interest to me is the upcoming x64 compatibility for Win-on-ARM, and whether or not Microsoft will license WoA for (legitimate and legal) use in VMs. Personally I see this as a more likely means of getting "Revit on Mac" without actually getting our much desired "Revit for Mac". It is worth noting Autodesk previously listed Windows running under Parallels Workstation as an officially supported environment.
For another path: if M$ can get x64 compatibility, and M$ licenses WoA to consumers for VMs (or licenses WoA for a polished off VirtualPC they bought from Connectix in the G4 and G5 era for the new Age Of ARM), and Revit "plays nice" under that degree of emulation, and the performance hit of running x64 Revit under ARM emulation on WoA in a VM on M1 Apple Silicon... Will i tbe possible to run Revit? Maybe. Well enough to use as one's daily driver? This is yet to be seen.
We have to realize what Apple did. They did not simply replace the X86 CPU with an Arm CPU. That would have done very little.
What they did was create a processor with heterogenious cores. In the M1 there are 8 ARM cores but even more other kinds of cores and itis there other cores that are contributing to the very impressive speed we see.
Not only did they include these specialized cores but they are all connected to a unified memory. Sharing RAM means nothing needs to be copied. It is the ver all design of the M1 that is fast, not just the Arm cores.
If at some point the PC Windows world moves to Arm, they will only be able to swap the CPU. Intel and AMD are in business to make "standard" chips not complete systems-on-a-chip like Apple is doing.
So a "port" of a CAD system to some future Arm-based Windows PC will do nothing to facilitate the port to an Apple M1. Apple software can be fast on the M1 because the chip designers and the software engineers work in the same building and can (and do) have meetings. If the software runs slow they can add special hardware to (say) a future M3 chip to replace the slow software with special-purpose hardware. Apple has already shown they are willing to add a half dozen of these special hardware cores to the M1. They do things like video rendering, encryption, neural networks, and signal processing all now moved outside of the CPUs.
It will be fun to look back in 10 years and see how the Windows PC world reacted. Apple's move could be the start of the end of the "standard architecture PC" that is built with off the shelf commodity parts. For decades we have been building PCs from just any mainboard we like, then we buy a CPU chip and then memory all from a different vendor. Those days may soon be gone because building a custom M1-like system gives better performance at a lower cost.
Today on a high-end PC we have some gigabytes of system RAM and some gigabytes of video RAM and a high-speed 16-lane PCIe bus to move data between the two memories. But on the M1 chip, the data never moves. Not moving is always faster than moving even if they make the bus faster. In the long run, the better method will win out.
When the days of standard architecture PC are gone so will be "standard PC software". AutoDesk and even Microsoft will need to sell software built to the special make and model computer so that it can take advantage of whatever special cores happen to be inside. Or it can ignore those cores and be written for "lowest common denominator" and suffer from a lack of performance
Apple is, I think 10 years ahead now. In a decade "everyone" will be running on custom hardware and the software vendors will have to deal with it or die.
So, the answer is "Yes" All software companies will be forced to offer native builds of their products on a wider range of hardware. As "Dell" software would run poorly on "HP" computers.
Autodesk would be insanely ignorant, if they don't imigrate to ARM! man the M1 out of the box kicks intel into space..
So if there is a chance, that we have a possible way to make REVIT work, then that is the chance.
Because the performance from REVIT in a real Projekt is right now not existing.
I know we all want BIM we all want 3D models and we all want to work together and become smarter.
M1 ist definitly the most exciting thing, since multicore, or since Steve came back to Apple! or the iPhone.
A beautiful Milestone in our history of Computer-Performance. And yes the Software need to adapt as well.
Autodesk, Adobe, and a lot of other Software-Giants.
Can't find what you're looking for? Ask the community or share your knowledge.