Apple Introduces High-Performance Mac M2 Processor
Computing
Speaker 1: Today, we begin our second generation of epi Silicon designed specifically for the Mac, the M CS chips deliver a scalability range of performance and capabilities. It began with M one's breakthrough set of features that transformed our most popular systems. M two starts the second generation of M C chips and goes beyond the remarkable features of M one, unlike others in the industry who significantly increase power to gain performance. Our approach is different. [00:00:30] We continue to have a relentless focus on power efficient performance. In other words, maximizing performance while minimizing power consumption combined with our unified memory architecture and custom technologies. This focus on power efficiency allows tu to bring even more performance and new capabilities to our most popular max MTU is built using and enhanced second generation five nanometer technology and features over 20 billion [00:01:00] transistors. That's 25% more than M one. We use those transistors to enhance every feature of the chip, starting with the memory controller, which delivers a hundred gigabyte per second of unified memory bandwidth, which is 50% more than M one for even greater performance.
Speaker 1: And because M one is so capable users on all sorts of demanding tasks. So for M two, we've enabled up to 24 gigabyte of unified memory to handle even larger and more [00:01:30] complex workloads. M two features our next generation CPU with advancements in both the performance and efficiency course, the performance cores are faster and feature larger cash, and the efficiency cores have been significantly enhanced for even greater performance gains together. They rip through CPU intensive tasks using very little power. When we look at the multi core CPU performance of M two compared to M one, it delivers 18% greater performance. [00:02:00] And compared to the latest 10 core piece of laptop chip the CPU and M two delivers nearly two times faster performance at the same power level, and to delivers the peak performance of the piece of chip while using a quarter of the power.
Speaker 1: Now let's compare M two to the latest 12 core piece of laptop chip, which needs dramatically more power to deliver an increase in performance. So it's in a thicker, hotter, more noisy system with less battery life. M two delivers nearly 90% of the peak performance [00:02:30] of the 12 core chip while using just a quarter of the power. This is what we mean by the power efficient performance of app and Silicon, and to also features our next generation GPU, which now has up to 10 course. That's two more than one combined with a larger cash and higher memory bandwidth. The Tanko GPU delivers a big boost in graphics performance, M two delivers up to 25% higher graphics performance at the same power level as M one and
Speaker 2: Up to 35% [00:03:00] higher performance at its max. And compared to the integrated graphics in the latest piece of laptop chip the GPU NM two delivers 2.3 times faster performance at the same power level, and two delivers the peak performance of the piece of chip using just one fifth of the power, this much higher performance per platform, M two enables systems to run cool and quiet with exceptional battery life. Now, one of the advantages of making our own Silicon is that we can quickly bring our newest [00:03:30] custom technologies across our SOCs. So M two also features our next generation secure enclave and neural engine. The neural engine in M two processes up to 15.8 trillion operations per second. That's over 40% more than M one. It also has our next generation media engine, including a higher bandwidth video decorder that supports AK H 2 64 and HVC video, and M two features our powerful pest video engine for hardware [00:04:00] accelerated encode and decode. So systems with M two will be able to play back multiple streams of 4k and eight key video. So that's M two, it starts our next generation of M CS chips with an 18% faster CPU, 35% faster GPU, 40% faster neural engine and 50% more memory bandwidth. It takes the amazing performance capabilities and EF of M one even further.
