To play this game we need obviously at least the M1 Ultra inside the MacBook Pro.From:
"On the Shadow of the Tomb Raider benchmark, the RTX was also a solid 30 frames per second faster. Now, this is Apple gaming, of course, so Tomb Raider was not a perfect or even particularly good experience: there was substantial, noticeable micro stutter at every resolution we tried. This is not at all a computer that anyone would buy for gaming. But it does emphasize that if you’re running a computing load that relies primarily on a heavy-duty GPU, the Mac Studio is probably not the best choice."
The bolded comment is the only point I am trying to make.
Apple charts and graphs don't do a great job of making this clear.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Has Apple lost their mind? March 8th Event Disappointment
- Thread starter NightOne
- Start date
- Sort by reaction score
Not sure if this is serious or not. I am literally playing this game currently on my M1 Mac mini at 1080p resolution and low settings.
Congrats.
But to buy Shadow of the Tomb Raider and play it at 1080p with low setting, this is like buying a Porsche with only the steering wheel.
You can probably play the game at 240p with much better settings!
From:
"On the Shadow of the Tomb Raider benchmark, the RTX was also a solid 30 frames per second faster. Now, this is Apple gaming, of course, so Tomb Raider was not a perfect or even particularly good experience: there was substantial, noticeable micro stutter at every resolution we tried. This is not at all a computer that anyone would buy for gaming. But it does emphasize that if you’re running a computing load that relies primarily on a heavy-duty GPU, the Mac Studio is probably not the best choice."
The bolded comment is the only point I am trying to make.
Apple charts and graphs don't do a great job of making this clear.
Apple Charts and graphs compare M1 Ultra vs RTX 3090 and M1 Max vs RTX 3060 Ti. I can pretty much confirm that the graphs are accurate as someone who owns M1 Max and RTX 3060 Ti desktop:
Affinity Benchmark (Raster):
M1 Max: 32288
RTX 3060 Ti: 16638
DaVinci Resolve Studio Export (DCI 4K HEVC):
M1 Max: 01:32
RTX 3060 Ti: 02:52
DaVinci Resolve Studio Export (6K BRAW):
M1 Max: 02:32
RTX 3060 Ti: 03:51
Blender BMW GPU Render:
M1 Max: 00:54
RTX 3060 Ti: 00:14
Blender BMW GPU Render:
M1 Max: 03:40
RTX 3060 Ti: 00:33
Shadow of thr Tomb Raider:
M1 Max: 83 FPS
RTX 3060 Ti: 128 FPS (158 with DLSS On)
GFXBench Metal / DirectX Aztec Ruins:
M1 Max: 311 FPS
RTX 3060 Ti: 310 FPS
Baldur's Gate 3:
M1 Max: 76 FPS (120 with FSR On)
RTX 3060 Ti: 74 FPS (110 with FSR On)
Photo/Video work is miles ahead on Apple Silicon. Even raw video work which does not utilize the encoders of M1 Max...
Blender is miles ahead on Nvidia as OptiX makes a huge difference. Apple and Blender are working on a similar support with Apple Neural Engine; however, I still think Nvidia will be on top for a while.
Baldur's Gate 3 comparison is pretty much on par with GFXBench Aztec Ruins, which shows the importance of a game being compatible with apple Silicon.
I do not even talk about the efficiency as the PC was much louder during all this use cases. If a tech reviewer brags about the product for photo/video/effects work for the entire half of the video and suddenly switches into GeekBench Compute and Shadow of the Tomb Raider comparisons on the second half, I would definetely question their judgement of "if you’re running a computing load that relies primarily on a heavy-duty GPU, the Mac Studio is probably not the best choice."
I have to play some new games at low settings with my GTX 1080 gaming PC..... Again, if you people are the spec chasers that need 4K, ULTRA, ray tracing, DLSS and more. Macs were NEVER and will NEVER BE for you.....EVER.....PERIOD. So why waste your time here? In the history of Macs it was never built for gaming compared to a dedicated $1,500 gaming PC.
So please, just get a gaming PC and be done with it. I have read these posts non stop since the 2010 Mac Pro for well over a decade.
Just like that $5,000 Dell I have referenced before, this Mac Studio and all Macs are geared toward professional use which means they basically suck for Gaming. That $5,000 Dell is not really better than my GTX 1080 when I tested it out. It has a Quadro GPU that barely matches a 3060 in some favorable tests. The M1 Ultra is cheaper than that Dell, but again, geared towards professional applications and not gaming.
Why aren't people breathing down those systems that Dell, HP, Lenovo makes? Oh right, because its "cool" to hate on Apple and it is very irritating.
Most people also forget that gaming cards are literally the cheapest junk out there - they are totally unsuitable for professional-level work. I only buy quadro/FirePro cards for my work machines (these run lab instrumentation and need to run 24x7 for 10-30 years) and the professional level quadro/FirePro cards are worth every penny. A single computer crash can (and has) cost me 10s of thousands of dollars and literally weeks of lost work.
Yep that is why we have a Dell workstation and not a beefy gaming PC because the Quadro is better than an off the shelf RTX 3080. There will NEVER be parity between gaming PC and professional systems. It can cover both areas to a certain extent, but when you get to the Pro of the Pro level, it does not. For example, I am able to render and work on my video projects on a gaming PC using Premiere Pro, but for 3D modeling I need that power of the Quadro.
(I have 7 different computers BTW each dedicated to a task or group of tasks that each does very well).
Is there a hardware difference? I though the drivers for the workstation cards were not handicapped for workstation loads like the "gaming" cards are.
There are differences sometimes. Quadro/Workstation RTX now has higher memory bus in some models, higher memory like the A600 has 48GB vs the 24GB in the 3090, slightly more CUDA cores.
• ECC on the professional (“Studio”) cards*
— Although, this does require somewhat slower VRAM: GDDR6 vs GDDR6X
• More VRAM (at each tier equivalent to the consumer models)
• Optimal quality processor chips for the professional cards* **
— For example, the RTX A6000 appears to utilize the full GA102 die (i.e. those that made it through production without any evident flaws)
• Different display outputs
• Lower power requirements
— From what I have seen, this is due to more stringent boost/turbo limits (which ensure greater consistency and reliability)
NVIDIA RTX 3090 vs. NVIDIA RTX A6000 vs. NVIDIA RTX A5000| GPU Benchmarks for AI/ML, deep learning 2024 | BIZON
In this article, we are comparing the best graphics cards for deep learning in 2023-2024: NVIDIA RTX 4090 vs RTX 6000, A100, H100 vs RTX 4090
* Presumably a factor of Nvidia willing to certify Studio/Quadro cards for use with professional applications, ensuring support and other coverage not guaranteed when using GeForce and other consumer-grade graphics cards.
** The RTX 3090 Ti is also claimed to house those flawless GA102 samples — assuming it ever launches
NVIDIA GeForce RTX 3090 Ti Graphics Card Specs, Performance, Price & Availability - Everything You Need To Know
NVIDIA GeForce RTX 3090 Ti 24 GB graphics card is coming out in a few weeks & here's all we know regarding its specs, price, & performance.
wccftech.com
Nvidia does offer a Studio driver compatible with GeForce cards.
NVIDIA Support
Last edited:
The consensus I’ve always found has been that Quadro cards are a waste of money and you can get similar performance far cheaper using consumer cards, at least in DCC apps (Maya, Houdini etc).
Which leads me to the question, where and what are you modeling where a Quadro makes any difference? Genuinely curious; the only benefit I can really see from a Quadro for modelling is the extra memory, but can’t see many situations where you’d really hit 48GB.
So the March 8th event did get me excited about one thing: getting a new Mac. Sadly this is not the one for me.
I did just make a decision and placed an order on Apple website today for:
drumroll …..
27” Mac Studio Display
M1 Mac Mini with 16GB RAM and 1TB SSD
The graphics performance is benchmarking at about half my current iMac but now I will be decoupled when a new M1 Max/M2 Mini comes out and I can upgrade again. Definitely one advantage over AIO setup.
I almost bought an i7 2020 refurb with 16B SSD and 5700XT card but I really do want to be on the new thing (Apple Silicon) and I’m happy with my decision.
Delivery times on display have slipped to the end of April so I will have to wait a bit.
I did just make a decision and placed an order on Apple website today for:
drumroll …..
27” Mac Studio Display
M1 Mac Mini with 16GB RAM and 1TB SSD
The graphics performance is benchmarking at about half my current iMac but now I will be decoupled when a new M1 Max/M2 Mini comes out and I can upgrade again. Definitely one advantage over AIO setup.
I almost bought an i7 2020 refurb with 16B SSD and 5700XT card but I really do want to be on the new thing (Apple Silicon) and I’m happy with my decision.
Delivery times on display have slipped to the end of April so I will have to wait a bit.
You hit it at the very end there. Its about the VRAM really. I needed more than the 3080 or 3090 has available. I primarly got my Mac Studio for my video projects with the extra video encoders and decoders, but will be trying other projects to see if we can migrate to that in the future with 64GB or 128GB of VRAM availability (obviously subtracting the application and OS memory but you get my point).
I think the inclusion of accelerators in the Apple silicon do have an impact on some of the traditional benchmarking methods. For example any hardware encoders will definitely increase performance vs software encoders.
You posted some good stuff so thank you for doing that.
There are definitely a lot of factors. I'm looking forward to see what comes in the future. (especially a Mac Mini update --- I would like to see something between the current mini and the studio -- M2?)
That would have made my decision easier but hopefully we have a newer mini version coming late 2022 or early 2023.
The more and more I think about it the March 8th event would have been the perfect time to roll out a more power Mac Mini. Although the marketing team might not have wanted that to happen.
I wouldn’t be at all surprised if it happens. Apple is still selling an Intel Mac mini, which indicates that there’s a hole in the Mac mini range still needing to be filled.
Unless perhaps the Intel Mac mini is still around to give customers that still need x86 for software support (but don’t need hardware expandability) a low cost Mac to meet their needs.
It’s a thing, sure, but Apple’s general purpose computational core also happens to be competitive generally, even without measures like performance per watt. I’m not sure how close other ARM licensees are to desktop class chips, since Apple’s mobile chips definitely have a significant competitive edge over, say, Qualcomm’s or especially MediaTek’s.
Performance per watt is something I personally do not care about. If it can get more performance by using more power then so be it. The Mac Studio plugs in the wall and doesn't run on battery so it really doesn't matter except for when picking a UPS.
Mac version doesn’t do raytracing. I don’t know what settings they tested the game on, but RTX was probably not enabled in the Nvidia benchmark.
Great, you don’t care…on a computer who’s primary target audience isn’t gamers. I don’t understand why this thread has just turned into a gaming thread. The studio is NOT designed to be a gaming machine.
In many gaming threads here and elsewhere, there has been an idea that if Apple wanted to make a gaming machine, it would put an M1 Max into a Mac Mini and sell it for $1500. The actual price is $2000 and the case is a bit bigger, but it's close enough. The games are still missing, but for all intents and purposes, the base model of the Mac Studio is indistinguishable from a gaming machine.
Except that Apple’s GPU’s are geared towards workloads. Yes it can game, but that’s not the point of the device and many here at MR seems to find that insulting somehow.
I never understood what is the factual basis for such a claim. Of course gaming is one of the main primary goals of the device. In fact, I would argue that G13 is better at rasterisation than at compute (Apple goes out of their way to make rasterisation and shading as efficient as possible). In fact, Apple is advising pro app developers to rewrite their compute workloads to take advantage of the gaming-oriented features such as tile shaders.