NotebookCHECK - Notebook Forum

All game developers should make their games run on Apple M, then gamers can peacefully let miners keep all the cards.

Apple advertise m1 max beating Rtx 3080 ? But in the end it cant even beat mobile Rtx 2060 from few years ago.

Sigh...Let's pretend we only do benchmarks.

If you work as a software DEV and build software locally, you're probably not impressed by the performance.

Imagine making powerful chips but then the game developer completely ignores your existence

>sailing past Core i9-12900HK in multi-core and reaching RTX 3070 levels of performance
M2 Pro/Max won't be out for a bit less than two year. By that time there will be Meteor Lake mobile and Lovelace mobile chips, and the Core i9-12900HK and RTX 3070 will be obsolete for any kind of benchmark

This is a completely ridiculous reasoning exercise.

"Nevertheless, it is not too difficult to hazard a guess that the M2 Max's GPU would be coming anywhere close to current high-end desktop offerings from AMD, Nvidia, and now Intel."

At the moment a M1Max GPU renders the Blender BMW27 scene in 41 seconds. A 3060 laptop renders it in 13 seconds.

You're kidding right? Why everyone tries to godify apple and this stupid chip

Quote from: toven on February 19, 2022, 16:50:40
All game developers should make their games run on Apple M, then gamers can peacefully let miners keep all the cards.

Apple has very low market share, especially away from USA. Then their higher end products cost similar or more than higher end Windows gaming laptops while providing less gaming value. Storage is limited and soldered, then their GPU is good for content creation while for gaming is way lower and lacks modern gaming features like raytracing or resolution "recovery" modes.

So there is close to no reason to release AAA games for Apple. The small market doesn't justify the cost and then people don't buy Apple products for gaming because there are no games for it ;)

Quotelacks modern gaming features like raytracing

Raytracing is important for content creation like Blender, Unreal and other GPU rander engines

dont just talk and make marketing jargon, just do some gaming trple A games, and then talk again  ;)

Annnnd STILL not interested in any Apple products.

The new MacBook's with M1, M1P, and M1X in-general are superior to competing laptops when it comes to the:
- build quality (screen, camera, trackpad, keyboard, shell, etc etc)
- software (macOS is much more optimised than Windows 10)
- hardware (much faster single-core, sustainable frequency, strong multicore, better GPU)

So they are definitely worth it, when looking on-paper or feeling it in the hand.

But none of that matters if you require a Program that runs on Windows (no macOS port). You could run it in a VM or Emulate it, but that creates bugs, kills compatibility, and defeats the purpose of your efficiency and performance.

So the MacBook's should not be seen through the lens of a benchmark, or how they might play games theoretically. They should be judged on what they can do, and that is the best experience for light use (surfing, office, etc) and a great use for content creators (specific macOS Applications).

Until you can either natively dualboot them into Windows, or the AAA-Gaming companies port their titles.... you shouldn't look at these as Gaming Laptops.

PS: currently Apple is happy where they are, they are the most profitable company when it comes to Video Games. Bigger than Xbox, PlayStation, Nintendo, and Steam combined. They get their money from iPods, iPads, and iPhone games because they take a 30% cut from all microtransactions. They're not interested in the next big game, which has massive costs and thin profit margins.

Apple are making excellent hardware for gaming and there is an Apple audience for this, Apple please take note, you need to fix Metal and engage better with the game development community to progress from where you are now. People want to play more than ported mobile games on their Macbook. These features need adding (Thanks Marlon for posting these on Apple developer forum)

Buffer device address:
This feature allows the application to query a 64-bit buffer device address value for a buffer. It is very useful for D3D12 emulation and for compatibility with Vulkan, e.g. to implement ray tracing on MoltenVK.

DrawIndirectCount:
This feature allows an application to source the number of draws for indirect drawing calls from a buffer. Also very useful in many gpu driven situations

Only 500000 resources per argument buffer
Metal has a limit of 500000 resources per argument buffer. To be equivalent to D3D12 Resource Binding Tear 2, you would need 1 million. This is also very important as so many DirectX12 game engines could be ported to Metal more easily.

Mesh shader / Task shader:
Two interesting new shader stages to optimize the rendering pipeline

Is this another story where someone just multiplied some numbers together and called it a day?

How can you get paid for articles when all you did was take the M1 performance numbers, multiply it by an arbitrary coefficient of BS and then proceed to tell us how crushing it might be even if you have a few concerns ever tendered hypothetical BS.

"At the moment a M1Max GPU renders the Blender BMW27 scene in 41 seconds. A 3060 laptop renders it in 13 seconds."

Nvidia Optix? Now try rendering an animation with it. It's useless except for the single image benchmark. Useless.

Quote"At the moment a M1Max GPU renders the Blender BMW27 scene in 41 seconds. A 3060 laptop renders it in 13 seconds."

Nvidia Optix? Now try rendering an animation with it. It's useless except for the single image benchmark. Useless.

Yes Nvidia Optix. It is not useless.
You telling that single image is only for benchmarks it shows you don't work professionally.

If you prefer only CUDA it is around 20 seconds.

It would be great to see M2 I & M2 MAXn near future , the performance projected and beating i9 -12900 HK is excellent from MAC prospective,  but one thing for sure it will still be lacking 100% on Gaming platform!

Geekbench doesn't accurately represent the performance of most applications on x86 and can be manipulated extremely easily especially under Linux.

Probably the only use case it produces accurate numbers for is application compilation as that is heavily memory bound like GB5.

Simply using clean Win7 or Linux distros gives you a good speed jump, let alone messing with kernel drivers & pre-emption.

Yes Nvidia Optix. It is not useless.
You telling that single image is only for benchmarks it shows you don't work professionally.

If you prefer only CUDA it is around 20 seconds."


As I said, OptiX is completely useless for any serious cgi work ie animations, usable only for fast previews, I give you that.
And yes, I do cgi/vfx professionally for various adds agencies.
The problem isn't OptiX denoiser itself, but the HYPE around it. Just stop it.

QuoteAnd yes, I do cgi/vfx professionally for various adds agencies.

Then you should know lots of work are stills not animation.

First, M2 Max isn't even coming this year. It will be launched in 2h2023 next year. With next year's perspective, it will be competing with raptor lake mobile, or may be even meteor lake and Zen4 mobile from AMD. So 1800 single core will put apple a generation behind competition. And 14k multi core will be nothing against AMD if 16 core zen4 mobile comes out at CES 2023. And AMD also very efficient now. That 6900hs beats M1 Max under 35 watt in cinebench r23, and 12 hrs of youtube playback with 72 whr battery.