Apple M4 officially presented – the new processor will use not only refreshed lithography, but also Mesh Shaders technology

Today, an Apple conference was held, which focused on presenting the new generation of iPad Pro and iPad Air tablets, as well as the Apple Pencil Pro stylus. During the presentation of iPad Pro (material about them will also appear soon), the manufacturer surprised quite a lot. We expected that the 2024 iPad Pros would be equipped with Apple M3 ARM systems, meanwhile the company decided to present the fourth generation of its Apple Silicon systems. Ultimately, the iPad Pro will be the first device powered by the Apple M4. What do we already know about the new processor?

Without any special announcements, Apple presented the M4 processor. Interestingly, it will first be used in this year’s iPad Pro tablets.

Apple MacBook Air laptop test with M3 processor. Quite efficient and elegant, but is it worth the price?

We have not yet learned all the details about the Apple M4 processor itself, at least in the context of dry technical specifications. The manufacturer confirmed the use of the second generation of the 3 nm technological process from TSMC, which is expected to result in better energy efficiency. The number of transistors has been increased again, although the increase is not as large as in previous generations. The Apple M3 had 25 billion transistors, while the Apple M4 has 28 billion. Apple has increased the number of cores, so that the basic M4 will now offer 10 cores, including 4 Performance cores and 6 Efficiency cores. Their timing remains a secret for now. The performance of M4 in iPad Pro compared to M2 in the previous generation of the tablet is to be higher by a maximum of 50%.

Apple M3 Apple M4 Technological process TSMC N3 (1st generation) TSMC N3 (2nd generation) Transistors 25 billion 28 billion Number of cores 8C/8T 10C/10T CPU configuration 4 P-Core + 4 E-Core 4 P-Core + 6 E-Core GPU Cores 10 Cores 10 Cores Memory Bandwidth 100 GB/s 120 GB/s Neural Engine 16 Cores, 16 TOPS 16 Cores, 38 TOPS GPU Features Hardware Ray Tracing
Dynamic Caching Hardware Ray Tracing
Hardware support for Mesh Shaders
Dynamic Caching Supports Tandem OLED No Yes

Apple M2 Ultra and Apple M4 processors can be used to build servers that power artificial intelligence

It is not known how much RAM the iPad Pro tablets are equipped with (Apple is not boasting about it yet), but we already know from the specifications that the bandwidth of the unified memory has been increased from 100 to 120 GB/s. Performance in AI calculations has been significantly improved. Previously it was 16 TOPS with 16 Neural Engine cores. Now the number of cores remains the same, but the power itself is 38 TOPS. At the same time, it is still less than for Qualcomm Snapdragon X Elite systems (45 TOPS) or the upcoming AMD Strix Point processors (from 50 to 60 TOPS), so in this respect there may be interesting clashes between these companies.

Related Articles:  AMD Zen4 low-power Ryzen7 7840U first appeared: 28W will kill 45W Zen3+ | T Kebang

The graphics system has been even further expanded with new functionalities. The number of cores is still ten, but now we can count not only on hardware acceleration of Ray Tracing, but also Mesh Shaders. The latter is responsible for the more extensive geometry of objects. There is also Dynamic Caching, which allows dynamic allocation of a specific amount of local memory in real time. Thanks to this, only as much memory is used for a given task as is actually needed. Apple M4 is also a new Display Engine that supports “tandem OLED”, i.e. a screen that uses the accumulated light of two OLED panels. Thanks to the support of such a display, the OLED parameters in the new iPad Pro tablets are so promising. The company declares a brightness of 1,000 nits in SDR, 1,000 nits in HDR on full screen and a peak brightness of 1,600 nits. The Apple M4 processor also supports ProMotion technology in the range of 10 to 120 Hz.

Source: Apple

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.