First Tests: Are the Apple M1 Pro and M1 Max Good for Gaming?

For decades, Apple has used third-party graphics processors in many of its laptops and desktops, from the humble Nvidia GeForce 9400M in early MacBooks to the muscular AMD Radeon Pro 5700 XT in the latest 27-inch iMac. 

The newest MacBook Pro laptops, by contrast, use Apple’s M1 Pro and M1 Max, custom-designed SoCs that combine CPU and GPU duties and share a pool of memory, to handle graphics output and acceleration. Seeing how well they perform on CPU-intensive workflows when we reviewed the 14-inch and 16-inch MacBook Pros got us thinking: Are these seriously powerful new processors equally adept at running graphics-intensive AAA games? 

It’s a fraught question. Apple makes it clear that its flagship MacBook Pro laptops are optimized for heavy-duty content creation, not gaming. And most game developers have overlooked macOS as a serious gaming platform, focusing instead on consoles and Windows. This chicken-and-egg situation means that even if the GPU portions of the M1 Pro and M1 Max are theoretically as capable as (or better than) the AMD Radeon or integrated Intel silicon they replace in past MacBooks, you still don’t have an extensive game selection to take advantage of.

Hitman (2016)

Still, a gamer can dream. So we ran a handful of graphics-intensive games (and a few less-demanding titles) released in the past few years on both the M1 Pro-equipped 14-inch MacBook Pro and the M1 Max-equipped 16-inch MacBook Pro to see what they are capable of. 

---

M1 Game Testing: Picking the Titles

Many of the AAA titles compatible with Macs come from British developer Feral Interactive, which partners with studios to bring their creations to platforms other than Windows. Among the most graphics-intensive Mac titles in our benchmark suite are the Tomb Raider series, from which we ran the built-in benchmarks in Rise of the Tomb Raider and Shadow of the Tomb Raider. These are older games, to be sure.

We also tested out a Total War game (Total War: Warhammer II), the Hitman reboot from 2016, and Sid Meier’s Civilization VI. The last game has far lesser demands on a computer’s graphics power than the rest of the lot, so we used its built-in benchmark for simulating how long it takes the AI to complete a turn, for a different perspective on game performance. All games were run at a full-HD-equivalent resolution (1,920 by 1,200 pixels) because the two new MacBook Pros have differing native display resolutions. (Testing at each laptop’s native resolution would have rendered the scores non-comparable.)

We also ran three graphics simulations: the cutting-edge 3DMark (specifically the Wild Life graphics simulation, recently released and compatible with macOS and Windows) and GFXBench Pro, as well as the older Unigine Valley test. These render scenes of richly detailed gaming-style graphics as a means to compare relative performance across devices with different operating systems. GFXBench renders the scenes without displaying them on the screen, and 3DMark renders them in a window, to account for resolution differences automatically, while we ran the Valley test at the same 1,920 by 1,200 pixels that we ran the games at.

(Photo: Tom Brant)

We also ran a subset of these games and synthetic tests on a GeForce RTX 3070-equipped gaming laptop we are in the process of reviewing: the Razer Blade 15 Advanced Edition. This Core i7-based system represents a state-of-the-art high-end Windows gaming laptop for comparison. Note that because the 1,920 by 1,200 resolution was not available on this Windows machine, we ran the games at a close equivalent of 1080p (1,920 by 1,080). The slightly fewer pixels being pushed on the Razer machine (the Apple machines are pushing around 11% more at 1,920 by 1,200) will give it a modest advantage.

Before we get to the results, some words about the actual graphics hardware in the new MacBook Pro laptops. The M1 Pro and M1 Max processors both have high-end and low-end configurations, and that affects the number of GPU cores. Both of our test units have the upgraded configurations of each CPU. The M1 Pro in the 14-inch laptop has 16 graphics cores and 32GB of memory shared between the CPU and GPU, while the M1 Max in the 16-inch MacBook Pro has 32 graphics cores and 64GB of shared memory.

With twice the number of graphics cores and memory in our sample M1 Max machine, it stands to reason that however good these two chips might be at rendering games, the M1 Max would be the better of the two by far. As it turns out, that isn’t always the case. 

---

Game Testing: A Look at M1 Pro and M1 Max Frame Rates

Without further ado, here are the results of our tests. Use the drop-down menu to cycle through all of the games and simulations as we explain them. Let’s start with the real-world games…

First up: the 2016 edition of Hitman. From the get-go, it’s clear that the new MacBook Pros are both theoretically capable of enjoyable experiences even when running truly taxing games. Sure, Hitman is an older title, but there still aren’t many non-gaming laptops made in 2021 that can average more than 100 frames per second (fps) on its in-game benchmark, and fewer still that can maintain that performance when you turn the detail settings up to Ultra. Surprisingly, the Macs run neck-and-neck with the Razer RTX 3070 system at Ultra!

The Hitman results also show that there’s something else besides the GPU power limiting performance, since the M1 Pro and the M1 Max have wildly different GPU capabilities but perform within a few frames per second of each other on this game, and indeed, lowering the graphical preset to Low didn’t increase frame rates by much.

The next title, Rise of the Tomb Raider, is even more demanding than Hitman. Unlike Hitman, this Lara Croft epic shows a large performance gap between the M1 Pro and M1 Max, more pronounced at Very High. Performance on AAA games always varies from title to title, even on Windows laptops, but the M1 chips have a few additional wild cards at play. For one, they’re so new that the programmers of these older games couldn’t have had their unique strengths and weaknesses in mind while coding. They also require an emulation layer known as Rosetta 2 to handle the CPU portion of the game’s processing demands. These games were originally designed to run on Intel CPUs, not the ARM architecture of the M1 Pro or M1 Max. The RTX 3070 rang up a near-tie with the M1 Max at Very High, again an impressive result for the 16-inch Apple MacBook Pro.

Shadow of the Tomb Raider

The big gap between the M1 Pro and M1 Max is still there when running the even-more-demanding Shadow of the Tomb Raider. In the case of this newer game, the 14-inch MacBook Pro equipped with the M1 Pro dips below the 60fps floor that many gamers demand on the Highest detail setting. The M1 Max stays comfortably above it at 84fps, however. (Remember, these are averages, and minimum frame rates dip into the 30fps range at times.) The RTX 3070 system held a clear lead, but the M1 Max MacBook Pro kept up gamely.

Total War: Warhammer II

Our next game, Total War: Warhammer II, is a bit less demanding than the preceding three. It also appears adept at harnessing the power of M1 silicon. On the Low detail setting, the M1 Max manages to put out an average of 176fps, compared with just 107fps for the M1 Pro. Clearly the additional graphics cores are put to good use here. 

The final title, Sid Meier’s Civilization VI, is a brief detour into the kind of performance Mac gamers can expect when the game isn’t graphically demanding at all, but rather uses the CPU to greater effect. Here, we’re looking at how long it takes the processor to complete a turn when you’re playing against the in-game AI computer. It’s a CPU-intensive task, and with CPU-core count the same between the M1 Pro and M1 Max (each has a total of 10 CPU cores), performance is equal. 

---

More Insights: Synthetic Graphics Testing

If you can’t tell by now, the answer to whether or not the M1 Pro and M1 Max are good for gaming is: “It depends on the game.” That’s a significant triumph for Mac silicon, since it’s essentially the same answer to whether or not an Nvidia or AMD GPU in a Windows gaming laptop is good for gaming. But it still doesn’t mean you should rush out to buy a $4,000 M1 Max laptop as your main gaming rig.

The clearest reason why not (other than the dearth of available games to play) is that Apple’s silicon is new enough that most games simply can’t use it to its full potential. A good illustration of this is the differences in the results of the GFXBench and Valley simulations. Let’s take a look at some synthetic graphics tests; again, hit the tabs, arrows, or dropdowns to see them all…

GFXBench and 3DMark Wild Life are the only applications mentioned in this story that run natively on M1 silicon, and the results are astounding. Breaking 300fps on the demanding GFXBench Aztec Ruins test places the M1 Max far above the M1 Pro, and on equal footing with the latest GeForce RTX laptop processors from Nvidia. Ditto for the 3DMark Wild Life trial.

But on the older, non-optimized Valley test, which is also extremely demanding, both M1 processors are clumped together down in the 80fps range. That’s still quite acceptable, of course, but it’s leaving a lot of performance on the table. For its part, Apple isn’t making any significant claims about the gaming performance of the M1 Pro and M1 Max the way it is for the chips’ ability to compile mountains of code or render multiple 8K video streams simultaneously. 

---

Surprising Gaming Contenders…for the Games You Can Find

Ultimately, the state of Mac gaming simply inches along with the introduction of M1 Max and M1 Pro, as good as the chips are. The hardware appears to be ready, but a critical mass of games that can take advantage of it simply doesn’t exist. On the other hand, if your preferred relaxation method at the end of a long day of content creation is revisiting classics like 2016’s Hitman, your new MacBook Pro will happily oblige.