Can your Mac handle Cyberpunk 2077? FPS benchmarks put M-series chips to the test

[midian182]

In brief: When Cyberpunk 2077 was released five years ago, nobody could have believed that it would one day be playable on MacBooks. But Cyberpunk 2077: Ultimate Edition finally arrived on Apple's devices last week. To see how well the game performs, a YouTube channel tested it on four Apple silicon-powered machines.

The Ultimate Edition of Cyberpunk 2077 launched on the Mac App Store as well as Steam, GOG.com and the Epic Games Store on July 17, combining the base game and the excellent Phantom Liberty expansion.

For the Mac release, CDPR spent 18 months rebuilding the REDengine pipeline to support Apple's Metal API. This involved optimizing the game for Apple's M-series chips and breaking down each frame into parallel tasks.

The studio says Cyberpunk 2077 is optimized for Apple silicon Mac models with 16GB or more of unified memory, though it recommends an M3 Pro with 18GB of unified memory for 1080p/60 FPS gameplay.

YouTube creator Andrew Tsai decided to put the FPS/RPG through its paces on four M-series Macs: an M4 MacBook Pro (16GB / 10-core GPU), an M3 Max MacBook Pro (48GB / 40-core GPU), an M1 Max MacBook Pro (32GB / 32-core GPU), and an M1 MacBook Air (8GB / 8-core GPU).

Not too surprisingly, the M3 Max with its 40-core GPU and 48GB of unified memory was the best performer, reaching 78 FPS at 1080p High Settings and MetalFX upscaling turned off. With MetalFX set to quality, that figure increased to 104 FPS.

While the unsupported 8GB M1 MacBook Air can run the game, you probably wouldn't want to play it. It manages just over 12 FPS with MetalFX enabled and set to quality – there was little point in testing it without upscaling.

The M1 Max MacBook Pro, with four times more RAM and GPU cores than the base M1 MacBook Air, does fare much better. It managed 70 to 90 FPS during indoor sections of the game and 60 – 70 FPS in open-world combat. The built-in benchmark test showed 51.87 FPS – with MetalFX enabled and with it disabled.

The M4 MacBook Pro with its 10-core GPU reached 26 FPS without upscaling. MetalFX pushed the figure to just under 39 FPS when enabled.

Tsai also compared Cyberpunk 2077's native Mac version against the Windows version running via the CrossOver compatibility layer. Somewhat surprisingly, the native port offered only around a 3% improvement.

Tsai said that he expects these figures will improve going forward as more optimizations are implemented.

Permalink to story:

Can your Mac handle Cyberpunk 2077? FPS benchmarks put M-series chips to the test

 

[techstrike]

Now CP2077 runs on Nintendo Switch 2... who cares if it runs on Mac?

 

[Theinsanegamer]

It's interesting that the performance improvement over parallels was only 3%.

Goes to show that all the claims of how x86 was so complicated and would be left in the dust by big ARM were greatly exaggerated.

Now CP2077 runs on Nintendo switch 2... who cares if it runs on Mac?

The millions of people who own a mac and might want to play the game?

What a silly comment.

 

[dangh]

Now CP2077 runs on Nintendo Switch 2... who cares if it runs on Mac?

I could say cp2077 runs on Steam Deck, who cares about Nintendo switch 2?

Hope you get it.

 

[Rudyduu]

I wonder if Apple subsidized CDPR to rebuild their game to run on MacOS and Mac hardware. I can't image they thought it would be worth the development time otherwise.

 

[techstrike]

It's interesting that the performance improvement over parallels was only 3%.

Goes to show that all the claims of how x86 was so complicated and would be left in the dust by big ARM were greatly exaggerated.

The millions of people who own a mac and might want to play the game?

What a silly comment.

I'm sure that mac users didn't wait this game to come to their platform, as they could buy a console for that reason ... it would cost them cheaper ... the funny part is that Mac silicon was hyped and in raw performance seems to be nothing special.

 

[techstrike]

I could say cp2077 runs on Steam Deck, who cares about Nintendo switch 2?

Hope you get it.

I believe NS2 is doing better than SteamDeck, running natively ofcourse.

 

[Billybobjoey]

They uploaded the Mac version onto Steam and GOG as well, so decided to throw it on my M1 Pro Macbook Pro to see how it would fare. I could get 60fps without frame gen as long as everthing was turned to low, but I decided to go for 40fps with everything set to medium and textures on high (no ray tracing because duh). It was an alright experience, aside from the amount of heat that was being dumped out. If I had the choice between playing it on a PS4/XB1 or my macbook, I'd probably choose the macbook. Obviously PS5/XBS XS and actual higher end PC's blow it out of the water, but for what it is I was pleasantly surprised.

 

[Theinsanegamer]

I'm sure that mac users didn't wait this game to come to their platform, as they could buy a console for that reason ... it would cost them cheaper ... the funny part is that Mac silicon was hyped and in raw performance seems to be nothing special.

Buying a console and the game will "cost them cheaper" than just....buying the game? Must be that nu math they teach in schools now.

I'm sure there is enough of a userbase that would like to buy the game on their mac to justify the development time, and I bet CDPR known more about that then you do.

I wonder if Apple subsidized CDPR to rebuild their game to run on MacOS and Mac hardware. I can't image they thought it would be worth the development time otherwise.

I highly doubt it.

The game is already finished, the effort to port it was likely a small dev team. We're not talking about a $100 million budget here.

There's also a lot of mac users out there. Mac sales have been doing great the last 3-4 years.

 

[dangh]

I believe NS2 is doing better than SteamDeck, running natively ofcourse.

Honestly, just a bit, and if I can choose between bit better battery life on the go or mods, netflix and youtube I'd rather go with steam deck.

But that was just written to address 'why someone plays on mac' stuff. I think ns2 is Inferior to pc handhelds but people have different needs and that's cool.

 

[axiomatic13]

It's interesting that the performance improvement over parallels was only 3%.

Goes to show that all the claims of how x86 was so complicated and would be left in the dust by big ARM were greatly exaggerated.

The millions of people who own a mac and might want to play the game?

What a silly comment.

Is your only purpose here to dunk on people? That's all you seem to do? Are you ever civil?

 

[Tom Yum]

To add a comparison data point, a 890m (Ryzen 7 370 - Strix Point) runs CP2077 at around 50fps at 1080p Low at ~50W SOC power draw, or ~30fps at 20W SOC (CPU + GPU) power draw, without FSR employed. Obviously High settings are more intensive than Low settings, so it isn't a direct comparison.

 

[Theinsanegamer]

Is your only purpose here to dunk on people? That's all you seem to do? Are you ever civil?

Did you come into the comments section just to complain?

I can be civil. Plenty of examples spread throughout this website. I also dunk on people wearing "dunk on my hot take" hats.

I'm sorry if this does not fit into your pure serious view of the internet. You are free to ignore me : )

To add a comparison data point, a 890m (Ryzen 7 370 - Strix Point) runs CP2077 at around 50fps at 1080p Low at ~50W SOC power draw, or ~30fps at 20W SOC (CPU + GPU) power draw, without FSR employed. Obviously High settings are more intensive than Low settings, so it isn't a direct comparison.

Sadly we do not have the M series power draw for comparison. If the M3 max was drawing ~50w, that would be a major achievement in their favor, if it's pulling 80+, less so.