News

Samsung Introduces Shinebolt HBM3E Memory Technology

Published

on


Advertisement

A global leader in cutting-edge memory technology, Samsung Electronics hosted its yearly Memory Tech Day today, displaying new memory products and industry-first breakthroughs to propel technical developments in a range of future applications, such as cloud computing, edge devices, and automotive technology. The keynote talk of Samsung Electronics’ President and Head of Memory Business, Jung-Bae Lee, focused on the company’s plans to innovate novel transistor architectures and materials in order to meet the difficulties of the hyperscale era.

Aiming to “reimagine memory,” the event, which drew in around 600 customers, partners, and industry insiders, gave Samsung officials an opportunity to elaborate on the company’s long-term strategies. LPDDR5X CAMM2, Detachable AutoSSD, and HBM3E Shinebolt are just a few of the new product developments that the business has unveiled. For instance, in order to enable greater single-chip capacities that potentially surpass 100 gigabits (Gb), Samsung is now developing novel 3D architectures for DRAM manufactured on a sub-10-nanometer (nm) technology.

Samsung Good Lock for Android 14: What to expect

Samsung has started mass-producing its 12nm-class DRAM in May 2023. Now, the company is developing its next-generation 11nm-class DRAM, which will have the greatest density in the market. Further advancements in NAND flash are being made to reduce cell sizes and improve channel hole etching methods, which will lead to the introduction of 1,000-layer vertical NAND (V-NAND). Samsung’s ninth-generation V-NAND, which uses a double-stack structure, is expected to be completed in time and offer the most layers in the market. In order to begin mass manufacturing next year, the business has obtained a working chip for the new V-NAND.

Revelation of the HBM3E ‘Shinebolt’:

  • Building on its success in launching the HBM2 market and enabling high-performance computing (HPC) with its first commercial HBM2, HBM3E Shinebolt DRAM was unveiled.
  • By lowering the total cost of ownership (TCO) and accelerating AI-model training and inference in the data center, it will enable next-generation AI applications.
  • Advertisement
  • With a 9.8 gigabits per second (Gbps) per pin speed, the HBM3E is able to attain transfer rates of 1.2 terabytes per second (TBps).
  • Samsung is already producing large quantities of its 8H and 12H HBM3 models, with product samples being shipped.
  • Also included were the first 32Gbps GDDR7 in the market, the petabyte-scale PBSSD that significantly increases server application storage capacity, and the 32Gb DDR5 DRAM with the largest capacity in the business.

Along with the industry’s first 7.5 Gbps LPDDR5X CAMM21, the company also demonstrated its 9.6 Gbps LPDDR5X DRAM, LLW2 DRAM designed specifically for on-device artificial intelligence, next-generation Universal Flash Storage (UFS), and the high-capacity Quad-Level Cell (OLC) SSD BM9C1 for PCs.

Samsung’s Exynos chips: Worth the trade-off for longer software support?

Samsung unveiled its Detachable AutoSSD, a virtual storage solution that enables data access from a single SSD to many SoCs. With a 4 TB capacity, the detachable AutoSSD may provide sequential read speeds of up to 6,500 megabytes per second (MBps). The SSD’s removable physical factor facilitates upgrades and modifications for both car makers and customers. Also available are automotive memory options with smaller package sizes, namely LPDDR5X and high-bandwidth GDDR7.

Technology that makes technology sustainable

Samsung’s semiconductor division will aim to achieve this goal by working with stakeholders across the semiconductor value chain, including customers and partners. The business intends to secure ultra-low-power memory technologies that can lower power consumption in PCs, mobile devices, and data centers. It also seeks to lessen its carbon impact by employing recycled materials in portable SSD goods. Samsung’s next-generation technologies, such as the PBSSD, will enable server systems to use less energy while maximizing rack capacity and space efficiency.

Advertisement

Exit mobile version