Hi. I have a Macbook M1 Max 32 GPU Core 32GB RAM. I use it for mobile development and works great for that purpose. A week ago I came up with an Idea for a game, I’ve read about game engines and the limitations of Unreal on MacOS and need some clarifications.
Will I have many assets not working ? Can I convert them by first installing them on a Windows virtual machine?
What systems doesn’t work on M1?
Is M1 Max a good machine to do an AA FPS game in Unreal?
Do I really need Lumen and Nanite to have nice graphics?
Again. I’m a noob. But this machine cost me 3000 euros. If I sell it for a windows machine I’ll have to develop for a living on an iPhone small screen. Since build for apple devices can be done only from apple devices.
For the most part you won’t have any issues with assets. Nanite assets can be converted to normal SMs in their details in editor. There are some plugins, and some lesser features that aren’t available either, but in most cases I wouldn’t worry too much here.
Most systems will work for M1 macs, but two key systems do not. Nanite does not work at all for M1 macs due to 64 bit atomics, and Lumen causes some artifacting from reports I’ve seen. This thread has some users discussing compatabilities with Lumen and Nanite (long read though) and even a user that almost got Nanite working on M1 macs.
You should be able to produce a AA quality FPS on your M1 just be wary to test builds often, and issues that arise during packaging, especially those that stem from Xcode issues can be hard to diagnose in larger projects.
Not at all! Nanite is exceptionally good at keeping performance for high poly geometry, but for many game ready assets, it’s entirely unnecessary and won’t provide a massive performance boost. For example, one project I’m on currently has absolutely no nanite geometry, because swapping over for low to mid poly assets provides little or no gains. Lumen is amazing at making nice lighting utilizing GI and emissives, but it’s by no means required for a game to look good.
Overall, I know many developers that even prefer to work from Macs, many older than your M1. Mac support is getting better over time as well, while some features aren’t able to be backported due to some constraints, the goal in the future was to have as much parity between OS’s as possible.
I would have gotten it working, if I didn’t pick up Robert Freitas’s Nanomedicine halfway through and find a passion for Molecular Nanotechnology. Kind of my fault for not following through. The reason for working on Nanite was a search for a purpose, after exiting the AI industry that disowned my hardware. I finally found that purpose with nanotech.
For the record, your exact setup is what I used to create an industry-first rendering algorithm with software ray tracing. Inspired by Nanite’s ability to render billions of polygons, I attempted to render billions of atoms. It has since grown into a much larger CAD framework, but I routinely push out ray traced renders that only Apple silicon can execute (YT playlist here). Reversed the paradigm of Mac users being treated second class.
An algorithm is an algorithm, and it can be implemented multiple ways. The path of least resistance is if it exists in hardware. However, I made a big point about doing 64-bit math on a 1-bit nano-sized microprocessor. It is possible, the question is how severe the tradeoff is in performance. For Nanite, I estimate it’s a small, but reasonable constant factor. The bigger issue is how much time the Epic devs are willing to invest into implementing this. They have every right to not support M1, with the Epic-Apple lawsuit. In engineering you have to make tradeoffs with a limited person-hours budget.
Epic added 64-bit atomics for M2 after I reverse engineered the architecture with the help of this community. I made some uarch documentation explaining what exactly is different between M1 and M2, and how Nanite can be done on both processors, in different ways.
Hey there @anjusinghal! Welcome to the community! There are no official plans announcement to backport for M1, however It’s absolutely technically possible, but would require a talented engineer on the level of Phil to pull it off. It’s recommended that if you remain in the MacOS ecosystem, to expect M2 to receive far closer feature parity going forward.