Live data from Hacker News

An Update on Apple M1/M2 GPU Drivers

lwn.net

221–230 of 291 posts

Re: An Update on Apple M1/M2 GPU Drivers

#221
post #189

Earlier quoted context omitted.

Using a laptop with 8gb RAM for a decade is an exercise in frustration

Only if you insist on running software that needs more RAM, in which case you shouldn't have bought it.

Apple should not have sold devices with 8GB or less back in 2018. Them doing it in 2024 is a sign that they think their users are idiots.

Re: An Update on Apple M1/M2 GPU Drivers

#222

Earlier quoted context omitted.

Thank you. This guy gets it. Path tracing is the gold standard for computer graphics. It is a physically based rendering model that is based on how lighting actually works. There are degrees of path tracing of varying quality, but there is nothing else that is better from a visual quality and accuracy standpoint. Your modern AAA title does a massive amount of impressive hacks to get rasterization into the uncanny val…

This is all built on the, frankly, nonsensical assumption that photorealism is the , let alone a , primary goal.

Not really. Tiny Glade is a recent ray-traced title that has no goal of photorealism whatsoever, and it looks gorgeous with RT lighting and shadows: https://youtu.be/QAUSBxxgIbQ

You can scale these same principles to even less photorealistic games like the Borderlands series or a Grand Theft Auto game. Ray tracing is less about photorealism (although it is a potent side effect) and more about creating dynamic lighting conditions for an interactive scene. By simulating the behavior of light, you get realistic lighting - photoreal or not a lot of games rely on SSR and shadowmaps that can be replaced with realtime RT.

Re: An Update on Apple M1/M2 GPU Drivers

#223

Earlier quoted context omitted.

Yes, MacBooks are generally more expensive to repair, but they also tend to not need repairs. It’s quite normal to hear from people who are using their 7+ year old MacBook without any issues and are still perfectly happy with it. I myself still use my 2018 MacBook Pro as my main device. When considering longevity, I will agree that Thinkpads are probably the only device that can compete with MacBooks. But there are t…

> MacBooks are going to be lighter Not going let Macs have this one, my X1 carbon is considerably lighter than a MBA. But generally agreeing. My last X1C lasted 8 years and still works perfectly, I just wanted an upgrade. My new one is even lighter than my old one. I opt for the FHD w/o touch screen and the second best processor to balance battery life and performance. Definitely not getting 18hrs battery life but 8-…

I admit I was assuming they would be heavier, I didn't consider the X1 Carbon. When I think of Thinkpads I still picture the traditional IBM-style Thinkpad. A quick look at the specs shows the 14" X1 Carbon at 2.42lbs, 13" MacBook Air at 2.7 lbs, and a 14" Thinkpad E series at 3.17 lbs.

Re: An Update on Apple M1/M2 GPU Drivers

#224

Earlier quoted context omitted.

Yes, MacBooks are generally more expensive to repair, but they also tend to not need repairs. It’s quite normal to hear from people who are using their 7+ year old MacBook without any issues and are still perfectly happy with it. I myself still use my 2018 MacBook Pro as my main device. When considering longevity, I will agree that Thinkpads are probably the only device that can compete with MacBooks. But there are t…

MacBooks are some of the heaviest laptps on the market. The Air is the "light" one at 2.7 lbs 13" and 3.3 lbs for the 15" For reference, there are several 13" laptops on the market that are around 2 - 2.5 lbs and 15"+ that are less than 3 lbs

Do any of those lighter laptops match the battery life and performance of the MacBook Air, while also being completely silent? I suppose I should have been more specific and stated I don't believe there are any laptops that can match the MacBook in all categories while being lighter.

Re: An Update on Apple M1/M2 GPU Drivers

#225

Earlier quoted context omitted.

Thank you. This guy gets it. Path tracing is the gold standard for computer graphics. It is a physically based rendering model that is based on how lighting actually works. There are degrees of path tracing of varying quality, but there is nothing else that is better from a visual quality and accuracy standpoint. Your modern AAA title does a massive amount of impressive hacks to get rasterization into the uncanny val…

Think of the 'impressive hacks' used in realtime 3D rendering as compression techniques to reduce bandwidth (and this goes all the way back to 8-bit home computers with their tricky video memory encoding, this was essentially hardware image compression to reduce memory bandwidth). Why waste computing power on 'physically correct' algorithms, when the 'cheap hacks' can do so much more in less time, while producing nea…

Because a lot of the "cheap hacks" aren't even cheap. Famously, using real-time SSR to simulate ray-traced reflections can end up slower than ray tracing them from the jump, the same goes for high-res shadowmaps. Ambient occlusion is a lighting pass you can skip entirely if you render shadows the right way with RT from the start. If you keep stacking these hacks until you have every feature that a RT scene does, you're probably taking longer to render each frame than a globally illuminated scene would.

Accelerating ray tracing in-hardware is a literal no-brainer unless you deliberately want to ostracize game developers and animators on Mac. I understand the reactionary "But I don't care about dynamic shadows!" opinion, but there's practically no opportunity cost here. If you want to use traditional lighting techniques, you can still render it on an RT-enabled GPU. You just also have the option of not wanting to pull out your fingernails when rendering a preview in Blender.

Re: An Update on Apple M1/M2 GPU Drivers

#226

Earlier quoted context omitted.

Using raytracing on a very recent, high end game that has been designed for raytracing has an extremely noticeable positive improvement on visual quality. Wukong is a perfect example. Unfortunately most people’s experience with raytracing is turning it on for a game that was not designed for it, but it was added through a patch, which results in worse lighting. Why? Because the rasterized image includes baked-in glob…

The first game I encountered with RayTracing support was World of Warcraft "Classic"'s Burning Crusade launch a couple years back. WoW Burning Crusade launched in * 2006* originally. The "Classic" re-release of the game uses the modern engine but with the original game art assets and content. Does it do anything in the 'modern' WoW game? Probably! In Classic though all it did was tank my framerate. Since then I also…

What's perhaps not obvious is that if you take a 2006 game and "add" raytracing, there's no relevant information available for the raytracer. Whatever original light sources were used to bake in global illumination does not exist in the game assets used for rasterization. In the best case the original art assets can be used to create raytracing scenes, but it's not as straightforward as it might seem, and most conversions are pretty sloppy.

If you want to see what raytracing can do, I'd only look at very, very recent UE5 titles that are designed for raytracing from the ground-up.

Re: An Update on Apple M1/M2 GPU Drivers

#227

I was going going to say she should work for Valve to help get Steam working on Linux on Macs, but it seems she already does? [1] [1] https://en.wikipedia.org/wiki/Alyssa_Rosenzweig#Career

Nice; that raises some interesting questions as to what Valve is planning here. Steam/Proton on macs would make a lot of sense for them hard as it may be. People booting linux on their macs to play games would really annoy Apple probably.

> People booting linux on their macs to play games would really annoy Apple probably.

You dont have to boot Linux to play PC games on Mac.

Apple already provides the tools you need to build a Mac native equivalent to Proton.

There are several options built using those tools, both paid {Crossover) and free/open source (Whisky).

Whisky combines the same ooen source Wine project that Proton leverages with Apple's Rosetta x86 emulation and Apple's DirectX emulation layer (part of the Game Porting Toolkit) into a single easy to use tool.

https://getwhisky.app/

Re: An Update on Apple M1/M2 GPU Drivers

#228

Since bringing modern OpenGL and Vulkan onto Apple Silicon is impossible without an emulation layer anyway, could, theoretically, a native Metal API for Linux be created? Or is Metal too ingrained in macOS SDKs? MoltenVK is attempting to solve the same issues Alyssa was talking about in her talk [1, the last comment on the issue is hers] [1] https://github.com/KhronosGroup/MoltenVK/issues/1524

Nothing is barring Apple from supporting Vulkan natively on MacOS. This is essentially the closing statement of Alyssa Rosenzweig´s talk. With Apple knowledge of internal documents they are the best positioned to produce an even better low level implementation. At this point the main blockroad is the opinionated point that Metal porting is the only official supported way to go. If Valve pull up a witch-crafted way to…

> If Valve pull up a witch-crafted way to run AAA games on Mac without Apple support that would be an interesting landscape. And maybe would force Apple to re-consider their approach if they don't want to be cornered on their own platform...

Right, except that Game Porting Toolkit and D3DMetal was an exact response to this scenario. Whether it's the right approach, time will tell, but Apple definitely already headed this one off at the pass.

Re: An Update on Apple M1/M2 GPU Drivers

#229
Alyssa's solution to the 4KB vs. 16KB page size discrepancy by running everything in a virtual machine feels like both a clever hack and a potential performance bottleneck. It makes me wonder about the long-term implications of such workarounds. Are we reaching a point where the complexity of bridging these gaps outweighs the benefits, especially when dealing with proprietary hardware designed to be a closed ecosystem?

This also touches on a broader question about the future of open-source efforts on platforms that are inherently restrictive. While it's inspiring to see games like Control running at 45fps on an M1 MAX with open-source drivers, it begs the question: Should the community continue to invest significant resources into making closed systems more open, or should efforts be redirected toward promoting more open hardware standards?

Apple's approach to hardware design warrants criticism. By creating GPUs with limitations that hinder standard functionalities like tessellation shaders and using non-standard page sizes, Apple places unnecessary obstacles in the path of developers. This closed ecosystem not only complicates the work for open-source contributors but also stifles innovation that could benefit all users.

Re: An Update on Apple M1/M2 GPU Drivers

#230

Earlier quoted context omitted.

Think of the 'impressive hacks' used in realtime 3D rendering as compression techniques to reduce bandwidth (and this goes all the way back to 8-bit home computers with their tricky video memory encoding, this was essentially hardware image compression to reduce memory bandwidth). Why waste computing power on 'physically correct' algorithms, when the 'cheap hacks' can do so much more in less time, while producing nea…

Because a lot of the "cheap hacks" aren't even cheap. Famously, using real-time SSR to simulate ray-traced reflections can end up slower than ray tracing them from the jump, the same goes for high-res shadowmaps. Ambient occlusion is a lighting pass you can skip entirely if you render shadows the right way with RT from the start. If you keep stacking these hacks until you have every feature that a RT scene does, you'…

Yes, for some specific problems, raytracing definitely makes a lot of sense, but rasterization and 'geometry compression' via triangles also makes a lot of sense for other situations. I think the best approach will always be a hybrid approach instead of a 100% pure raytracing pipeline.

The other question is how much of the raytracing features in 3D APIs need to be implemented in fixed-function hardware units, or whether this is just another area where the pendulum will swing back and forth between hardware and software (running on the GPU of course).

But then maybe in 10 years we'll have 'signed distance field units' in GPUs and nobody talks about raytracing anymore ;)

Post reply on HN