Apple releases M1 Ultra chip: 114 billion transistors, 20-core CPU+64-core GPU
With the release of the Mac Studio, Apple also introduced its latest M1 series chip: the M1 Ultra. This is a huge leap forward for Apple and its M1 series of chips, through an innovative packaging architecture called UltraFusion, Apple interconnects two M1 Max chips, creating SoCs with unprecedented levels of performance and functionality, giving Mac Studio incredible computing power.
Made up of 114 billion transistors, the M1 Ultra is the largest ever on a personal computer chip. Apple has configured the M1 Ultra with up to 128GB of high-bandwidth (up to 800 GB/s), low-latency unified memory, plus the highest 20-core CPU (16 performance cores + 4 energy efficiency cores), 64-core GPU, and 32-core NPU. It provides developers with unparalleled performance and undertakes 3D rendering tasks that were difficult to complete in the past. Like the previous M1 Max chip, each performance core has 192KB of instruction cache, 128KB of data cache, and a total of 48MB of L2 cache. Each energy-efficient core has 128KB of instruction cache, 64KB of data cache, and a total of 8MB of L2 cache.
After Apple launched the M1 Max chip last year, some people have studied its structure, the M1 Max was found to have an undiscovered interconnect bus that could support stacking through specific interposers and multi-chip packages, in theory. It is concluded that in the future, Apple may continue to expand based on the architecture of the M1 series of chips. The appearance of the M1 Ultra this time also confirms that this idea is correct.
To build the M1 Ultra, Apple used UltraFusion, a custom package architecture, to connect the two M1 Max chips. This method solves the problems of increased latency, reduced bandwidth and increased power consumption caused by dual-socket motherboard configuration with dual CPUs, and two chips can be recognized as one CPU, allowing developers to take full advantage of its performance. The M1 Max’s interposer has more than 10,000 signal pins, providing a low-latency bandwidth of 2.5 TB/s between processors, more than four times the bandwidth of the best multi-die interconnect technologies available today.
M1 Ultra has a 64-core GPU — 8x the size of M1 — delivering faster performance than even the highest-end PC GPU available while using 200 fewer watts of power.
The M1 Ultra’s media engine supports hardware-accelerated H.264, H.265, ProRes, ProRes RAW, with two video decoding engines, four video encoding engines, and four ProRes encoding and decoding engines. It also doubles the performance of the M1 Max, delivering higher throughput for ProRes video encoding and decoding, allowing Mac Studio to play up to 18 streams of 8K ProRes 422 video. The M1 Ultra also integrates custom technologies such as a display engine capable of driving multiple displays and an integrated Thunderbolt 4 controller. It also features best-in-class security features, including Apple’s latest Secure Enclave, hardware-verified secure boot, and runtime anti-exploitation technology.
“M1 Ultra is another game-changer for Apple silicon that once again will shock the PC industry. By connecting two M1 Max die with our UltraFusion packaging architecture, we’re able to scale Apple silicon to unprecedented new heights,” said Johny Srouji, Apple’s senior vice president of Hardware Technologies. “With its powerful CPU, massive GPU, incredible Neural Engine, ProRes hardware acceleration, and huge amount of unified memory, M1 Ultra completes the M1 family as the world’s most powerful and capable chip for a personal computer.”
