The march of Apple silicon will not be slowing down. To energy the brand new high-end Mac mini and 14-inch and 16-inch MacBook Professional techniques, Apple has unveiled the all-new M2 Professional and M2 Max processors, taking all of the enhancements of the A15 and A16 processors and making use of them to the Mac chips, leading to a major increase in efficiency over the M1 line. Apple claims the M2 Professional and M2 Max chips carry 20 p.c higher CPU efficiency and 30 p.c sooner GPU efficiency than the M1 Professional and M1 Max. Additionally they embody the newer Neural Engine, which is as much as 40 p.c sooner than that within the M1 sequence.
As spectacular as these claims are, they’re additionally unsurprising and a bit disappointing. Among the extra thrilling rumors about extra superior manufacturing, new GPUs, or different enhancements have confirmed unfaithful, and the M2 Professional and M2 Max look like precisely what was anticipated: the M2 scaled up with extra CPU and GPU cores and a wider reminiscence bus. In truth, the M1 Professional and M1 Max chips had extra to differentiate them from the M1 than the M2 Professional and M2 Max do when in comparison with the M2.
However nonetheless, there are some fascinating issues to find out about these chips earlier than you resolve to purchase one of many new Macs. Listed here are 5 key traits value noting.
Similar CPU core design and efficiency as M2
When Apple launched the M2, it claimed an 18 p.c enchancment in CPU efficiency versus the M1. Now it’s saying that the M2 Professional and M2 Max have 20 p.c higher CPU efficiency than the M1 Professional and M1 Max, and leaked benchmarks appear to again up these claims. In different phrases, you’ll be able to anticipate the identical total architectural enhancements and efficiency uplift because the M2 delivered in comparison with the M1.
Apple
A few of this merely comes right down to core depend. The M2 Professional and M2 Max have eight high-performance cores identical to the M1 Professional/Max, however the variety of effectivity cores goes up from two to 4. So the M2 Professional and M2 Max have 12-core CPUs whereas the M1 Professional/Max have 10-core CPUs.
GPU beneficial properties are higher, however there’s no nice architectural shift
Apple says the M2 Professional/Max GPUs are 30 p.c sooner than the M1 Professional/Max, and leaked Metallic benchmarks again up these claims. That’s unsurprising, contemplating the M2 was mentioned to have graphics efficiency that’s as much as 35 p.c higher than the M1.
That is probably as a result of addition of some extra graphics cores and a few minor architectural tweaks. The M2 Professional has as much as 19 graphics cores (vs. 16 within the M1 Professional) and the M2 Max has as much as 38 graphics cores (vs. 32 within the M1 Max).
These are uncommon numbers—core counts are sometimes even in quantity, and infrequently powers of two, so it’s probably that the design is for 20/40 GPU cores with room for one defective core to be disabled, bettering manufacturing yields on these massive chips. Nonetheless, that’s a virtually 20 p.c enhance in core depend, so with some improved caching and better clock speeds, you will get to 30 p.c.
Apple
Should you had been hoping for an all-new GPU from Apple with fashionable options like ray-tracing acceleration, which GPUs from AMD and Nvidia have had for a number of years already, you’re going to be dissatisfied.
Reminiscence bandwidth is similar even with extra RAM
Reminiscence bandwidth on the M2 Professional is 200 GB/sec, identical to the M1 Professional. The M2 Max doubles that to 400 GB/sec. So we’re virtually actually nonetheless a 256-bit LPDDR5 reminiscence bus on the Professional and 512-bit LPDDR5 on the Max. Apple improved caching within the A15, and people architectural enhancements probably carry over to the M2 Professional and M2 Max.
Whereas the M2 Professional maxes out at 32GB of RAM (the identical because the M1 Professional), the M2 Max can go as much as 96GB–50 p.c greater than the 64GB restrict of the M1 Max. However there’s a catch—the 96GB RAM choice is proscribed to the model of the M2 Max with 38 GPU cores. The cheaper choice with 30 GPU cores tops out at 64GB.
They nonetheless don’t have AV1 help
With the M1 Professional had a brand new video encoder and decoder that added {hardware} acceleration for the ProRes format, and the M1 Max doubled the variety of video encoders to 2. After some confusion, that’s nonetheless the case with the M2 Max, but it surely helps all the identical codecs as earlier than: H.264, HEVC, ProRes, and ProRes RAW. Which means help for the newer AV1 codec continues to be nowhere to be discovered, regardless of being out there in at present’s GPUs from Nvidia, AMD, and Intel. Apple is falling behind right here.
Apple
We couldn’t assist however discover that each one of Apple’s advertising and marketing benchmarks round video manufacturing, claiming enhancements of 30 p.c or so, are for actions like shade grading or 3D results rendering, sometimes restricted by GPU efficiency. Apple didn’t present benchmarks exhibiting an enchancment in video encoding pace, so it is going to be fascinating to look at the primary opinions of the M2 Max to see if the video encoders are measurably sooner.
They’re nonetheless all 5nm
There was plenty of hypothesis about when Apple will begin utilizing TSMC’s new 3nm manufacturing course of. Apple is claimed to be the primary main producer to take action, and a few of the rumors mentioned the brand new M2 Professional and Max chips could be the place Apple makes the leap.
It appears like we’ll be ready for an M2 Extremely later within the yr, or presumably the A17 that might be launched with the iPhone 16 this fall. Apple says the M2 Professional and Max are made utilizing “second-generation 5nm expertise.” It’s probably that this is similar course of that Apple considerably erroneously referred to as “4nm” in its iPhone 15 presentation in September 2022. TSMC calls its N4 node an “enhanced” 5nm course of, and Apple appears to be correcting its language right here.