Huawei's Mate 10 smartphone is powered by the AI-specific Kirin 970 chips, built using a 10nm process atTSMC featuring an octa-core ARM Cortex CPU, a Mali-G72 12-core GPU and a neural processing unit (NPU) designed specifically for a mobile device, according to the vendor.
The NPU, combined with Huawei's HiAI mobile computing platform, enables the Kirin 970 to deliver 25x better performance and 50x greater energy efficiency for AI-related tasks, compared to four Cortex-A73 cores, Huawei claimed.
The Mate 10 comes with a 5.9-inch LCD Quad HD screen with a 2,560 by1,440 resolution, 4GB RAM and 64GB internal storage that can be expanded up to 256GB via a microSD card. It features dual 4G SIM support and dual VoLTE connections.
Samsung HKMG DDR5
Samsung Electronics has expanded its DDR5 DRAM memory portfolio with a 512GB DDR5 module...
Photo: Company
Nvidia GeForce RTX 30 series GPUs
Nvidia's GeForce RTX 30 series GPUs are powered by the company's Ampere architecture. The...
Photo: Company
Apple HomePod mini
Apple's HomePod mini is the newest addition to the HomePod family. At just 3.3 inches tall,...
Photo: Company
Apple 13-inch MacBook Pro with Magic Keyboard
Apple has updated the 13-inch MacBook Pro with the new Magic Keyboard for an improved typing...
Photo: Company
Apple iPad Pros
Apple's new iPad Pros comes with the latest A12Z Bionic chip, an ultra-wide camera, studio-quality...
Photo: Company
As Microsoft transitions from a software giant to a cloud leader, with its cloud business now accounting...
Mobile, telecom, computing trends
CSP in-house development of ASIC accelerators
Google TPUs will see a share of over 70% in the in-house developed cloud ASIC accelerator market in 2024; an all-optical network...
AI chip market outlook 2023-2028: Insights from demand and supply perspectives
The growing demand for AI computational power is accelerating advancements in hardware and chip technology, necessitating innovation...
Automotive CIS tech development, 2024
The popularization of autonomous driving is boosting demand for automotive CIS with LFM and HDR being mainstream development...