Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Are they good enough to play games?. I mean, how do they compare with a PC with a graphics card?

Terribly. Previously I posted an article showing the M2 is better than Intel integrated graphics, but not significantly better. And even some of the worst PC graphics cards destroy the graphics performance of the M2. As for a good, modern PC graphics card? Forget about it, not even in the same universe.
 
  • Like
Reactions: SilverWalker
OLED can’t match the brightness and have other issues like burn in. The screens on the MacBook pros are outstanding. OLED wouldn’t be an upgrade to these.
Apple fans were saying these exact things about OLED on the phones until Apple started using them. According to rumors iPads are next. Eventually Macs will follow. On average, Apple is some 10 years behind Samsung in OLED adoption.
 
  • Like
Reactions: SFjohn and M3gatron
Is it not true that Apple counts or considers GPU “cores” differently than AMD and Nvidia count/consider GPU cores?

I used to think it was comparable to CPU cores — like getting ±8/16± whole GPUs!

(I do know about ALUs, but, still, I don’t think Apple is comparing apples to apples.)

What’s the true story?
There is no right answer because GPUs are a hierarchy of "compute units" clustered together to share more and more items. At the lowest level you might have a set of "compute units" that share a scheduler (ie which instructions to send to the "compute units") and register file, at a higher level this "core" might share an L2, but at a lower level this core might be split into four "sub-cores" each with a separate L1 cache. Or other permutations!

These details matter because they determine things like how fast one set of threads can share data or synchronization with another set of threads. Beyond that is the issue of "how fast for WHAT?" Graphics performance depends not just on how many FLOPs, but also on things like how rapidly textures can be read, or on specialized hardware like geometry shaders; whereas AI performance depends on things like what types of numbers (FP16? BF16? 4bit integer?) are supported.

You can get some idea of how the various designs compare by looking at:
https://github.com/philipturner/metal-benchmarks
which (at the very top of the page, click on the arrow!) gives some numbers for Apple vs recent AMD and nV designs. It's clear that there has been some variation in how large "cores" are and how they are balanced over the past 8 years or so, but that all three vendors have now converged onto something very similar.
 
OLED can’t match the brightness and have other issues like burn in. The screens on the MacBook pros are outstanding. OLED wouldn’t be an upgrade to these.

Absolutely incorrect, on multiple levels. You need to go do some research on the significant gains made in the PC/OLED space, including QD-OLED. As for IPS screens being on-par with OLED? Plainly -- plainly -- you've never done a side-by-side comparison. OLED colors, HDR performance, infinite contrast, true pure blacks, and the insanely fast pixel response times (fastest pixel display tech on the market by far), just blow IPS out of the water.

Once you've used an OLED monitor with your computer -- you won't go back to IPS or VA.
 
Okay so performance and longevity wise, which of the these two same priced systems am I better off with:
  • Mac Studio - Apple M1 Max with 10-core CPU, 24-core GPU, 16-core Neural Engine, 32GB, 1TB SSD; or
  • Mac Mini - Apple M2 Pro with 12‑core CPU, 19-core GPU, 16‑core Neural Engine, 32GB, 1TB SSD
Wait for benchmarks. Probably the mini tho
 
Absolutely incorrect, on multiple levels. You need to go do some research on the significant gains made in the PC/OLED space, including QD-OLED. As for IPS screens being on-par with OLED? Plainly -- plainly -- you've never done a side-by-side comparison. OLED colors, HDR performance, infinite contrast, true pure blacks, and the insanely fast pixel response times (fastest pixel display tech on the market by far), just blow IPS out of the water.

Once you've used an OLED monitor with your computer -- you won't go back to IPS or VA.
Even QD-OLED are not as bright as the brightest miniLED. On that part he is correct. Also, fast pixel response time is good for gaming, but ruins 24p content like movies and prestige tv shows with horrific stutter.
 
I made a comment about this when the "expected roadmap" was shared earlier in Jan. Their roadmap makes zero sense, right now, studio as a product, is dead in the water.

Not because the studio is bad. Great machine for people that already bought it, but there is very little rationale for buying it right now.
Obviously it will be updated with m2 ultra eventually
 
Studio's relative "fatness" is apparently fat mostly to deal with Silicon heat dissipation. The Mac Mini case appears to have no changes from when it was all Intel in there. Presumably something has to give in throttling or fan noise. Else, if all very positive assumptions stick- that Mini will be as quiet as Studio and able to run PRO at full speed all of the time, a case for "Why Studio?" starts gaining more steam... at least until there's a Studio with M2 or M3.
The current Mini design is designed to accommodate an Intel i7-8700B with a 65W TDP, and cranking it up to maximum would cause the fans to go crazy.

The M1 Pro and Max have a TDP roughly half that (Apple doesn't give official TDP figures but third party sites have extrapolated). The Ultra is close to the TDP of the 8700B meaning it would begin to push the limits of the Mini's thermal design. Teardowns show the Ultra has a completely different heat sink and fan design versus the Max version of the Studio.

Thus the Studio's thickness is primarily to accommodate cooling for the Ultra and ensure it can always run at peak performance without the fans sounding like a jet engine. There shouldn't be any thermal constrictions of an M2 Pro inside of the Mini's enclosure just like there aren't many restrictions inside the MBP.
 
Okay so performance and longevity wise, which of the these two same priced systems am I better off with:
  • Mac Studio - Apple M1 Max with 10-core CPU, 24-core GPU, 16-core Neural Engine, 32GB, 1TB SSD; or
  • Mac Mini - Apple M2 Pro with 12‑core CPU, 19-core GPU, 16‑core Neural Engine, 32GB, 1TB SSD

CPU bound workflows, the Mac Mini with M2 Pro will be faster - even for software that cannot take advantage of multiple cores.
In terms of general work performance, there is probably not much difference between them in real world use.
The M1 Max supports 3 displays (I think) while the M2 Pro supports 2. So, unless you need three displays, the Mac Mini would win (in my opinion of course).
 
  • Like
Reactions: SFjohn and Tagbert
OLED colors, HDR performance, infinite contrast, true pure blacks, and the insanely fast pixel response times (fastest pixel display tech on the market by far), just blow IPS out of the water.
Note that while OLED has faster response times than LCD, what matters for motion blur is instead MPRT, and MPRT times for OLED and LCD are comparable:

 
Last edited:
Exciting news. But, can it run Crysis?

I know the reference, but in seriousness... I'm interested to see how it runs Resident Evil 8. I don't plan on getting or playing it... strictly curiosity since the M1 runs it very well.
 
That’s not a real question, right??? They’re two different classes of desktop, one with a massive heatsink, more ports, more bandwith…a processor revision wouldn’t really outweigh all of those other factors.
It is a real question. one the fortunately others have asked and helpful people on this forum have attempted to answer. You do realise that a new version of one class of desktop could outperform an older version of another. Not really that difficult to work out.
 
I'm hoping Apple Silicon will be a good contender for a gaming computer.
Not only that, I think Apple Silicon and an SDK that is friendlier to game developers would make the Mac a superior gaming platform.

Apple just doesn’t seem particularly interested in a respectable library for the Mac of the most current and popular games available on the major consoles and PC.

Their strained relationship with Nvidia, their lack of OS support for code that was written to run on the graphics hardware of the three major consoles and on the PC/PC graphics cards, and their uncompromising stance on the use of the Metal API only — even eschewing support for the “MoltenVK graphics and compute API” makes Apple come off as disdainful of the indie and AAA game developers. And the results are clear.

Apple used to have “Software Evangelists,” one of whom would approach/meet with game developers and essentially ask, “What do you need from us? How can we help? What difficulties do you face coding games to run on the Mac, and what changes to the platform would make your jobs easier? Would you like a team of Apple engineers to fly out to your studios and work directly with your programmers to get around any hurdles you may face writing for the Mac? Can we offer training? Need any money?”

So it wasn’t always the way it is now. Bungie’s original Halo was designed to come out on the Mac first and a near-complete version was first demonstrated to the world by Steve Jobs on stage at MacWorld in 1999. It caused waves in the press because the gameplay and graphics represented a paradigm shift in the gaming industry at the time.

Bill Gates was also impressed by the demo and quickly bought Bungie to get his hands on Halo — and promptly cancelled all versions of Halo for all other platforms and made it exclusive to the Xbox.

(This wasn’t Apple’s fault, but the point remains that this indie game studio, Bungie, chose to develop “Halo: Combat Evolved” for the Mac first — because it was such a powerful and attractive and exciting platform to write for at that time.)

ADDENDUM: Apple does have its own subscription gaming service, Apple Arcade, where almost every game must work across all Apple devices: iPhone, iPad, Apple TV, Mac — even iPod touch. The only issue here pertaining to the Mac is that games made to work on all Apple devices don’t get to stray in terms of compatibility with iOS to avail themselves of the more power and features of the Mac/macOS. Using Apple Arcade, it’s more the case that you’re playing an iOS game on a Mac.

Still, outside of the Apple Arcade Ecosystem, game developers are free to write games titles specifically for the Mac and macOS and all the additional power and versatility the Mac platform affords — they just aren’t.

The recent trend in gaming — resolution scaling — is all the rage with AMD‘s FSR upscaling method along with Nvidia’s DLSS and Intel’s XeSS all duking it out, while Apple adds to the melee with its own proprietary MetalFX.

It was interesting to read that Apple nearly included hardware-accelerated ray tracing in its latest generation of Silicon SIPs — by adding hundreds of tensor cores to Apple Silicon I would assume. Time will tell.
 
Last edited:
  • Like
Reactions: SFjohn
Indeed. These are pretty darn good increases given that the chips are on a very similar process as M1.

All signs have been pointing to the M3 as the one to look forward to for a while now. It'll be on the 3nm process and, if we believe the rumors, should have a good chance of having the new greatly upgraded GPU cores, maybe with ray tracing support.

Still, these M2 machines should be great for anyone who gets them. If I hadn't bought an M1 24" iMac a couple of years ago I'd be all over the M2 Pro Mini (of course what I really want is an Mx Pro 24" iMac... looks like that's not coming any time soon, although maybe with the M3?)

Anyway: exciting times for the Mac these days. :)
I'm more interested in the M3 processor, but do not understand where most of the improvements will be.

I understand the increase in performance and possible addition of ray tracing. I am more curious to know how the 3nm chips will improve battery life on the Macbook Air and Pro.
 
Significant improvement compared to the previous generation.
Indeed. I think people were disappointed by the MBA M2 speed increased, but Apple smartly released that with otherwise completely new hardware.

But now that the Pro got two more cores, and the GPU also got an amazing uplift, Apple nailed the pro releases, and is getting really good PR.
 
  • Like
Reactions: Swissfashion
Considering even the regular M1 could decently play Resident Evil Village, I guess it's good, and for a laptop it's awesome, and considering thermals too.

But if you look at the market in terms of performance, the Apple Silicon GPUs are quite a few years behind the competition, especially considering the price. And I haven't even talked about ray-tracing yet.
 
Last edited:
All of you praising OLED on Samsung.. but I find image on Iphone much better looking than on any Android phone.

Open the same app on iPhone and on the best Android phone, and it looks better to the eye on iPhone.
 
Okay so performance and longevity wise, which of the these two same priced systems am I better off with:
  • Mac Studio - Apple M1 Max with 10-core CPU, 24-core GPU, 16-core Neural Engine, 32GB, 1TB SSD; or
  • Mac Mini - Apple M2 Pro with 12‑core CPU, 19-core GPU, 16‑core Neural Engine, 32GB, 1TB SSD
The studio has twice the memory bandwidth as the mini, 400GBs vs 200GBs.
 
Eh not quite. Yes, in graphics performance it's getting there. But CPU performance on the Ultra M1 is 23,000 which is still more than a 50% increase from the Max M2.

Which means the Ultra M2 should approach 30,000 CPU and 120,000 GPU

(Geekbench scores)

Keep in mind the Ultra is bandwidth limited... there really isn't much improvement in the 64 s 48 cores. If Apple addresses that in the M2 Ultra it could be roughly 2x the speed of the Max.
 
Geekbench doesn't even flood the cores of discreet GPUs either. W6800X Duo 1 core gets around 140,000 Metal compute score and I never saw the usage go above 40% on that GPU core while being tested.

Compute apps that can tap multiple GPUs like Octane simply fly on Mac Pros with W6880X Duos.

Honestly not concerned about discrete GPU's. The issue is with the score discrepancies between M1 variants. The GPU cores across all variants of the M1 run at the same speed, so performance should actually scale linearly with only a minor loss for each additional core. (Apple has increased memory bandwidth of each variant to make sure those cores can be "fed" without any lag. On the Max, Anandtech tested the bandwidth, and was able to push data to all GPU cores with a sustained bandwidth rate of 240GB/s. So we know it's not a memory bandwidth issue.)

Geekbench Metal scores...
M1 8-core, ~10W TDP, 68GB/s: 20440
M1 Pro 16-core, ~30W TDP, 200GB/s: 39758 (+95%)
M1 Max 32-core, ~60W TDP, 400GB/s: 64708 (+317%)
M1 Ultra 64-core, ~120W TDP, 800GB/s: 94583 (+463%)

As expected, the first tier does in fact scale linearly; double the cores, double the performance. But as we get higher the scores drop way more than they actually should. The Ultra loses over 300% which is an insane drop in performance.

As a side note: Unfortunately until we get applications that are actually optimized for Apple's GPUs (Metal), we won't see the performance they're actually capable of.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.