Apple M5 Ultra and M6 chips are here
Apple today unveiled two new chipsets that are already being used in the new Mac mini M6 and the new Mac Studio M5 Max and M5 Ultra.
Apple M5 Ultra: New flagship
According to Apple, the M5 Ultra is the most powerful chip ever. It consists of two dual-die M5 Max chips connected using Apple’s next-generation UltraFusion technology. This marks the first time Apple has used a quad-die architecture in an M-series chip. UltraFusion increases the bandwidth between the dies to over 4.4 TB/s and the connection density by more than six times, allowing the four dies to function as a single processor.
The Apple M5 Ultra includes a CPU with up to 36 cores, including 12 super cores and 24 performance cores, offering up to 1.25 times the single-threaded performance and up to 1.3 times the multi-threaded performance of the M3 Ultra.
The Apple M5 Ultra’s GPU scales to up to 80 cores, with a neural accelerator integrated into each core for the first time in an Ultra chip. According to Apple, this delivers up to 4.3 times the peak performance for AI calculations compared to the M3 Ultra, as well as up to 1.8 times faster graphics performance than the previous generation chip.
The Apple M5 Ultra also features a 32-core Neural Engine and a more powerful Media Engine with twice as many video encode and decode blocks as the M5 Max, allowing users to play up to 33 streams of 8K ProRes 422 at 30 fps simultaneously.
The Apple M5 Ultra supports up to 512 GB of unified storage and a storage bandwidth of up to 1.2 TB/s – a 50% increase over the M3 Ultra. According to Apple, the additional storage and higher bandwidth enable the local storage of large datasets, the processing of higher tokens per second, and the execution of complex language models with hundreds of billions of parameters entirely on the device.
The Apple M5 Ultra is available exclusively in the new Mac Studio, which can be pre-ordered starting today and will be available from September 22nd. Configurations with 512 GB of RAM will be available at the end of October.
Apple M6: First 2nm chip is here
Apple also unveiled the M6 chip, the first Apple chip manufactured using a 2-nanometer process and featuring a dual 16-core Neural Engine. Apple states that the 2nm process allows for higher transistor density on a smaller chip, thus contributing to the chip’s overall performance and energy efficiency gains.
The M6 chip features a larger and more powerful 12-core CPU—two more cores than the M5—consisting of two supercores, four performance cores, and six efficiency cores. According to Apple, the chip offers the world’s fastest single-threaded performance and up to 1.2 times faster multi-threaded performance than the M5, as well as up to 2.4 times faster performance than the M1.
The Apple M6 also features a 12-core GPU, with two more cores than the M5, each with an integrated neural accelerator. According to Apple, this delivers nearly 30% higher peak GPU performance for AI compared to the M5 and more than eight times that of the M1, resulting in faster processing of on-device LLMs.
The GPU also receives updates to the shader core architecture, dynamic caching and hardware-accelerated ray tracing, as well as a 50% higher geometry rate for more complex graphics.
The Apple M6 features a dual 16-core Neural Engine, which surpasses the maximum processing power of previous generations by half. According to Apple, system frameworks can automatically utilize both engines simultaneously, allowing applications to run AI models faster.
The chip supports up to 32 GB of unified memory and up to 170 GB/s of unified memory bandwidth, which is a 10% increase over M5 and 2.5 times over M1.
The Apple M6 makes its debut in the new Mac mini, which will be released on Tuesday, September 22nd.
