Setting issues up for what’s actually to be an thrilling subsequent few months on the planet of CPUs and SoCs, Apple this morning has introduced their next-generation M-series chip, the M4. Introduced simply over six months after the M3 and the related 2023 Apple MacGuide household, the M4 goes to begin its life on a really totally different observe, launching alongside Apple’s latest iPad Pro tablets. With their latest chip, Apple is promising class-leading efficiency and energy effectivity as soon as once more, with a selected give attention to machine studying/AI efficiency.
The launch of the M4 comes as Apple’s compute product strains have change into a bit bifurcated. On the Mac facet of issues, all the current-generation MacBooks are primarily based on the M3 household of chips. On the opposite hand, the M3 by no means got here to the iPad household – and seemingly by no means will. Instead, the newest iPad Pro, launched in 2022, was an M2-based machine, and the newly-launched iPad Air for the mid-range market can be utilizing the M2. As a consequence, the M3 and M4 exist in their very own little worlds, a minimum of for the second.
Given the fast turn-around between the M3 and M4, we’ve not come out of Apple’s newest announcement anticipating a ton of adjustments from one technology to the subsequent. And certainly, particulars on the brand new M4 chip are considerably restricted out of the gate, particularly as Apple publishes fewer particulars on the {hardware} in its iPads on the whole. Coupled with that may be a give attention to evaluating like-for-like {hardware} – on this case, M4 iPads to M2 iPads – so info is thinner than I’d prefer to have. None the much less, right here’s the AnandTech rundown on what’s new with Apple’s newest M-series SoC.
Apple M-Series (Vanilla) SoCs | ||||
SoC | M4 | M3 | M2 | |
CPU Performance | 4-core | 4-core 16MB Shared L2 |
4-core (Avalanche) 16MB Shared L2 |
|
CPU Efficiency | 6-core | 4-core 4MB Shared L2 |
4-core (Blizzard) 4MB Shared L2 |
|
GPU | 10-Core Same Architecture as M3 |
10-Core New Architecture – Mesh Shaders & Ray Tracing |
10-Core 3.6 TFLOPS |
|
Display Controller | 2 Displays? | 2 Displays | 2 Displays | |
Neural Engine | 16-Core 38 TOPS (INT8?) |
16-Core 18 TOPS |
16-Core 15.eight TOPS |
|
Memory Controller |
LPDDR5X-7700 8x 16-bit CH 120GB/sec Total Bandwidth (Unified) |
LPDDR5-6400 8x 16-bit CH 100GB/sec Total Bandwidth (Unified) |
LPDDR5-6400 8x 16-bit CH 100GB/sec Total Bandwidth (Unified) |
|
Max Memory Capacity | 24GB? | 24GB | 24GB | |
Encode/ Decode |
8K H.264, H.265, ProRes, ProRes RAW, AV1 (Decode) |
8K H.264, H.265, ProRes, ProRes RAW, AV1 (Decode) |
8K H.264, H.265, ProRes, ProRes RAW |
|
USB | USB4/Thunderbolt 3 ? Ports |
USB4/Thunderbolt 3 2x Ports |
USB4/Thunderbolt 3 2x Ports |
|
Transistors | 28 Billion | 25 Billion | 20 Billion | |
Mfc. Process | TSMC N3E | TSMC N3B | TSMC N5P |
At a excessive degree, the M4 options some type of new CPU advanced (extra on that in a second), together with a GPU that appears to be largely lifted from the M3 – which itself was a brand new GPU structure. Of specific focus by Apple is the neural engine (NPU), which continues to be a 16-core design, however now presents 38 TOPS of efficiency. And reminiscence bandwidth has been elevated by 20% as effectively, serving to to maintain the extra highly effective chip fed.
One of the few issues we will infer with a excessive diploma of certainty is the manufacturing course of getting used right here. Apple’s description of a “second generation 3nm process” strains up completely in timing with TSMC’s second-generation 3nm course of, N3E. The enhanced model of their 3nm course of node is a little bit of a sidegrade to the N3B course of utilized by the M3 sequence of chips; N3E just isn’t fairly as dense as N3B, however…