M2 Ultra vs Nvidia RTX 4090

The M2 Ultra and the RTX 4090 represent flagship products from Apple and Nvidia, each bringing their best performance to the table. However, the approach from each company couldn’t be any more different.

Nvidia is what one considers the “old school” of thinking where you specialize in doing one thing and doing it very well, by any means necessary. This means bumping up the clock speed and power consumption to the limit that components might melt from the heat. Apple, on the other hand, prefers to take a holistic approach by identifying the workflow and building solutions that attack said workflow. And the result is the M2 Ultra where there are specialized cores to tackle specific workflow.

At the end of the day, the user wants results, and which gives the better results? The M2 Ultra in the Mac Studio / Mac Pro or that custom-built PC which is Nvidia RTX 4090.

Executive Summary:-

• M2 Ultra was launched by Apple during WWDC 2023 keynote address
• RTX 4090 was launched by Nvidia around October 2022 during the GPU Technology Conference event.
• Mac Studio (a complete computer unit) with M2 Ultra starts at $3,999 while the Nvidia RTX 4090 card alone starts from $1,700 to $2,000.
• As far as Apple Silicon goes, the M2 Ultra is Apple’s most powerful, most capable, and fastest Apple Silicon ever and features the best graphic cores in the entire Apple Silicon lineup. It is a marvel of engineering with a combination of specialized cores like Media Engine, Display Engine, AMX, and Neural Engine, The M2 Ultra is pretty fast doing certain workflows
• But in terms of raw power, the RTX 4090 is simply an order of magnitude more powerful than the M2 Ultra will ever be. The M2 Ultra couldn’t even beat Nvidia’s last generation of graphic cards the RTX 30-series in terms of raw power.
• Apple prefers to measure performance in their own workflow, and as far as the users in Apple’s ecosystem, the M2 Ultra is the best Apple Silicon ever.
• Apple’s M2 Ultra advantage is being far more power efficient. Apple’s Mac Studio which has the M2 Ultra uses one-third of the power required by the RTX 4080 graphic cards alone.

M2 Ultra

The UltraFusion in the M2 Ultra

The M2 Ultra was introduced in WWDC 2023. At the same time, Apple put the M2 Ultra in two Macs: the Mac Studio and the Mac Pro. With the update of the Mac Pro, Apple finally sealed Intel’s fate in Apple and Apple had built everything in-house, bringing one of the rare computer companies that is vertically integrated.

Like the M1 Ultra before it, the M2 Ultra is the combination of two M2 Max co-join together at the hip like Siamese twins using the UltraFusion technology. With UltraFusion technology, bandwidth between the two M2 Max chips can be as high as 2.5 TB/s, which is far higher than any competitor products like NVlink or even AMD Zen2 microarchitecture. The UltraFusion bandwidth is high enough that those two M2 Max chips can be considered as a single monolithic M2 Ultra chip by macOS.

Spec bump on the M2 Ultra includes higher memory support to 192 GB, faster processing cores, and more importantly in the context of this article: up to 76 graphic cores. Each of the graphic cores is an improvement over the previous generation M1 series and there’s more of them in M2 Ultra.

Apple is the “think different” company in that it approaches things differently than, say Microsoft in handling graphic cores. Microsoft for example gives frameworks like Direct X as a guide on how the graphic card is supposed to interface with the OS. Manufacturers basically have carte blanche on how to implement the drivers to comply with the framework. Apple on the other hand would like to be more involved in everything that goes with macOS.

Apple also prefers to put specialized cores based on their internal studies of how the machine will be used by their users. So most of the emphasis is on handling media files, doing AI tasks like pattern recognition, and being decent at gaming. Decent is the keyword here.

Apple also has upgraded the I/O capabilities of the M2 Ultra. Now you can hook up three 8K displays on the Mac Studio and are capable to run 4K displays at 240Hz, which is quite an achievement.

Nvidia RTX 4090

Nvidia is a company that specializes in computer graphics since the concept of computer graphics was in people’s imagination. Rise from fighting competitors, continuous improvement, and plain luck, Nvidia is constantly producing the world’s fastest and best graphic chips on the market. When thinking of PC: Master Race, one can attribute Nvidia to the moniker.

The RTX 4090 is the latest from Nvidia from a long line of super graphic cards that have dominated the market over decades. By now, Nvidia has got designing and building high-performance graphic cards down to a science. And it same goes for the RTX 40-series: more graphic cores, higher clock speed, and better microarchitecture. The result is the RTX 4090 which consumes around 450 watts (around 2-3 times the entire Mac Studio computer) is around 60% to 75% faster than the previous generation’s top-of-the-line RTX 3090 card.

What’s new for the latest RTX 40-series is Nvidia is using TSMC’s 4N manufacturing process. TSMC also makes Apple’s chip such as the M2 Ultra and A16 that is used for the iPhone. There’s also the 4th generation tensor core which helps AI workload and the 3rd generation ray-tracing engine (which Apple doesn’t have at all). The micro-architecture is named Ada Lovelace in honor of the world’s first computer programmer.

Is Nvidia’s path to dominance is sustainable? One might say not since the latest card has some teething issues like having the new connector melted. But this is not the first time Nvidia has faced some technical challenges, which they can overcome. In fact, with the explosion of AI, Nvidia is the latest company the join the $1 trillion club.

Performance

Performance is what you expect: M2 Ultra can’t best Nvidia’s weakest graphic card from the last generation, even in Apple’s bread and butter: video editing.

Conclusion

By now, you have heard two normal stories: Apple’s latest chip while being a marvel of engineering, graphics-wise can’t even beat the weakest of Nvidia’s last generation. And it holds true even in the M2 Ultra. Benchmark after benchmark, game after game, the M2 Ultra is not the last word in gaming (or video editing for that matter).

So what is M2 Ultra’s trump card against the RTX 4090? In a word: ecosystem. M2 Ultra is Apple’s flagship chip and no doubt is competitive (although does not best against Intel/Nvidia), but M2 Ultra is a cog of a larger, well-oiled machine: Apple’s ecosystem. In Apple’s world, the M2 Ultra is Apple’s ultimate chip and yes, that’s true if you compare it against other Apple Silicon chips. But Apple is still a long way from being the last word in graphic performance.

The great thing about tech is nothing is ever static, so Apple’s graphic performance will get better in the future and might even rival Nvidia RTX 4090. But the bad news is tech is never static. So Apple’s competitors like Nvidia are not standing still either.

Plug

Support this free website by visiting my Amazon affiliate links. Any purchase you make will give me a cut without any extra cost to you

Other Ecosystem Items

• Apple TV (Amazon) / Apple TV 4K (Amazon)
• AirTags (Amazon)
• HomePod Mini (Amazon)