Apple has introduced the M5 chip, its next big leap in performance and efficiency for MacBook Pro, iPad Pro, and Apple Vision Pro. Built using third-generation 3-nanometer technology, M5 improves nearly every part of Apple silicon with faster AI processing, a stronger GPU, and higher memory bandwidth. All devices powered by M5 are available for pre-order today.

Powerful GPU Built for AI and Graphics

The M5 features a new 10-core GPU with a Neural Accelerator in each core. This design lets the GPU handle AI workloads much faster, offering over four times the peak GPU compute performance compared to M4 and more than six times compared to M1. The GPU also brings third-generation ray tracing and improved shader cores, resulting in up to 45 percent higher graphics performance than M4 and up to 2.5 times faster than M1.

The GPU’s new dynamic caching and optimized architecture deliver smoother gameplay, faster rendering, and more realistic visuals in creative and 3D apps. On Apple Vision Pro, M5 helps render 10 percent more pixels on the micro-OLED displays, with refresh rates up to 120Hz for sharper detail and reduced motion blur.

Faster CPU and Neural Engine

The chip includes a 10-core CPU with four performance cores and six efficiency cores, giving up to 15 percent faster multithreaded performance over M4. It also features a faster 16-core Neural Engine that handles AI-powered tools like Apple Intelligence and spatial photo creation on Vision Pro with greater speed and efficiency. Developers using Apple’s frameworks such as Core ML and Metal 4 can take advantage of the Neural Accelerators for improved AI app performance.

Improved Memory and Energy Use

M5 has a unified memory bandwidth of 153GB/s, about 30 percent higher than M4 and more than double that of M1. It supports up to 32GB of unified memory, letting users run large AI models or creative apps like Final Cut Pro and Adobe Photoshop at the same time.

(Source)

The post Inside the Apple M5 Chip: Next-Gen GPU, Faster Neural Engine, and Improved Efficiency appeared first on Gizmochina.