Earlier this year, Apple debuted the M4 chip in the updated iPad Pros. Since then, it’s been with us for a while, but surprisingly, we still know very little about it mainly because it has only been in an iPad so far. However, we are just a few weeks away from seeing how the M4 series of chips will perform in a Mac. As of last month, production has started for both the M4 Pro and M4 Max chips, which makes this fall season incredibly exciting for tech enthusiasts. In this article, I’ll break down why that is by first taking a look at the evolution of Apple silicon over the past three generations.
Performance Comparison: M2, M3, and M4
To give some context, here are the Geekbench 6 multicore scores for the entire M2 lineup, the M3 lineup, and the M4 chip. Right off the bat, you’ll notice some key points. The M3 was a pretty significant leap forward, with the M3 Max scoring the same as the M2 Ultra. But what’s even more impressive is how the M4 chip scores. It’s matching the performance of both the M2 Pro and M3 Pro.
At this point, you might wonder: Why are the M2 Pro and M3 Pro scoring the same? Let’s rearrange the data and sort by score from low to high to get a clearer picture. When we add the core breakdowns, it all makes sense. The M2 Pro and M2 Max both have 8 performance cores and 4 efficiency cores, whereas the M3 Pro trades two of those performance cores for efficiency, leaving it with an even 6 performance and 6 efficiency cores. Here’s where the M4 chip becomes impressive it achieves the same score with 50% fewer cores than the M3 Pro and half the performance cores of the M2 Pro. And remember, this is all happening in an iPad, which uses passive cooling.
How Did Apple Make the M4 So Powerful?
So how did we get to this point? How did Apple make the M4 chip so powerful? To answer that, we’ll need to get a little technical.
The M4 chip runs on Apple’s second generation 3-nanometer process (N3E), which brings several improvements to performance. The efficiency cores now run at 2.88 GHz (up from 2.75 GHz), and the performance cores hit 4.5 GHz (up from 4.05 GHz). Apple also switched to LPDDR5-7500 RAM instead of LPDDR5X, which allows the M4 chip to achieve an enormous 120 GB per second bandwidth 15 GB per second faster than the M3. This change also helps reduce memory response time. Another exciting rumor is that the upcoming M4 Max will finally start with 16 GB of RAM, making this an exciting time for Apple silicon.
Apple’s Silicon Architecture: A Technical Dive
Now, let’s dig deeper into what makes the M4 chip different. Apple has once again widened the decoding unit, allowing it to decode 10 instructions per clock cycle compared to 9 on the M3 and 8 on the M1 and M2 chips. Additionally, Apple has improved the dispatch buffer on the floating point unit (FPU), allowing it to handle more operations in parallel and boosting efficiency.
But what does all this technical jargon mean? Let’s use a simple analogy. Think of the CPU as a factory that operates on a clock cycle. This clock cycle represents the time it takes to complete one full operation of fetching, decoding, executing, and storing data. The decoder, or “foreman” of the factory, has become smarter with the M4 chip, handling 10 instructions at once instead of the 9 and 8 of previous generations. After decoding, the instructions are sent to the dispatch buffer (local “managers”), which can now handle more entries at the same time, thus increasing both performance and efficiency.
These changes are actual silicon improvements, not just tweaks to run the chip faster, as was the case with the M2, which used more power for slightly higher performance. Instead, with the M4, Apple has increased both performance and efficiency without adding more cores or increasing power consumption.
The M4’s Impressive Jump in Single-Core Performance
The single-core score for the M4 is over 20% higher than the M3, marking the biggest jump in Apple silicon performance since the original M1 chip. This tells us that Apple, with the M4 generation, has used the new process and architecture to boost peak performance. They’re not focusing on making the chips consume less power but rather on producing more work. For example, the M4 iPad Pro scores 46% higher in multicore performance compared to the M2 iPad Pro an impressive leap.
The M4 even matches the performance of the M3 Pro and M2 Max. This leads to some exciting possibilities for the M4 Pro chip, which could potentially match the performance of the M3 Max or even get close to the M2 Ultra. The M4 Max, when released, could be an absolute powerhouse. The M3 Max already matches the M2 Ultra, despite having fewer performance and efficiency cores, and the M2 Ultra is effectively two M2 Max chips fused together.
What About the Ultra Chip?
Here’s where things get interesting. Apple broke with tradition in the M3 generation by removing the UltraFusion interconnect from the M3 Max, which meant there was no M3 Ultra. This suggests that Apple may have a different strategy in mind for the M4 generation. Could they be developing a new technology that allows them to create an Ultra chip without combining two Max chips? Time will tell.
Mark Gurman has reported that Apple is planning three main versions of the M4 series: the base M4, higher-end versions codenamed “Bravo” that will replace the M3 Pro and Max, and an M4 Ultra, codenamed “Hydra.” The M4 brings significant changes to the CPU architecture, but the GPU side remains relatively similar. The M4 GPU has about a 10% higher clock speed than the M3, which means it consumes more power and produces slightly better benchmarks. Apple seems to be using improved binning for the GPUs that have been in production for a while.
What to Expect at Apple’s October Event
With the October Mac event just around the corner, what can we expect? One major highlight is likely to be a redesigned Mac Mini, which hasn’t seen a design change in nearly 15 years. We’re also expecting M4 iMacs, and M4, M4 Pro, and M4 Max MacBook Pros. However, don’t expect updates for the Mac Studio, Mac Pro, or MacBook Air just yet those are likely waiting for Ultra chips or scheduled for next year.
For Mac Mini fans, this could be the event you’ve been waiting for, as Apple may leapfrog over the M3 and go straight to M4 and M4 Pro models. As for the MacBook Pro, this release coincides with the third anniversary of the current design, and Apple typically refreshes its designs every 4-5 years. So if a redesign is important to you, it might be worth waiting until the next generation.
Whatever Apple does next with this generation of silicon, it’s bound to be exciting. Stay tuned for more updates on the M4 series.
