Samsung unveils next

发布时间：2023-12-07 21:22:12 来源：grand 888 slot 作者：컴퓨터

Samsung Electronics unveiled its newest chip technologies and products at its annual global event in San Jose, California, Friday, eyeing a key role in providing hyperscale artificial intelligence models.

Under the theme of “Memory Reimagined,” about 600 people, including technicians, analysts and industry officials, attended this year’s Memory Tech Day to look into Samsung’s technologies in areas such as cloud, edge devices and automobiles.

The tech giant unveiled its latest memory solutions, including the new fifth-generation HBM3e memory called Shinebolt; LPDDR package-based module LPDDR5X CAMM2 solutions and detachable AutoSSD that can be used through storage virtualization.

The Shinebolt provides high performance of up to 9.8 gigabits per second per data input and output pin, which can process over 1.2 terabytes per second. Samsung implemented high-level stacking by optimizing non-conductive film technology by filling gaps between chips.

Samsung’s industry-first 7.5Gbps LPDDR5X CAMM2 also captured attendees’ attention at Friday’s event. The chip threatens to be a game-changer in the next-generation PC and laptop DRAM market.

To achieve the No. 1 position in the memory market by 2025, it also introduced its Detachable AutoSSD, which supports continuous read speeds of up to 6,500 megabytes per second and provides 4TB of capacity.

“The era of hyperscale AI models is a point where opportunities for technological innovation and growth meet together. It will be a time of challenge as well as greater leaps for the industry players,” Lee Jung-bae, executive vice president and head of US memory at Samsung Electronics said.

“We will continue to lead the memory market by providing expanded solutions that overcome limits through infinite imagination and bold challenges, alongside close cooperation with customers and partners,” he said, stressing the firm will introduce new structures and materials to overcome the challenges faced in the era of hyperscale AI models.

相关文章