Rambus announces the completion of the HBM3 memory subsystem topping at 8.4 Gbps of transfer speed

Rambus Incorporated, founded in 1990, is an American technology company that designs, develops and licenses chip interface technologies and architectures that are used in digital electronics products. The company is well known for inventing RDRAM and for its intellectual property-based litigation following the introduction of DDR-SDRAM memory.

Rambus, a renowned chip manufacturer has just announced that it has completed the development of its advanced HBM3 memory subsystem. According to the manufacturer, this memory subsystem can hit transfer speeds of up to 8.4 Gbps.

Currently, in the market, we have the HBM2E which boasts to be the fastest memory option available and offer transfer rates of up to 3.2 Gbps. However, the HBM3 aims to offer more than double the speed of the current market standard offering an insane 8.4 Gbps transfer speed resulting in higher bandwidth.

According to sources, a single HBM2E package can peak out at 460 GB/s bandwidth however an HBM3 will offer up to 1.075 TB/s bandwidth which is 2x the total bandwidth jump.

Rambus had this to say about the technology:

The memory bandwidth requirements of AI/ML training are insatiable with leading-edge training models now surpassing billions of parameters,” said Soo Kyoum Kim, associate vice president, Memory Semiconductors at IDC. “The Rambus HBM3-ready memory subsystem raises the bar for performance enabling state-of-the-art AI/ML and HPC applications.”

Rambus achieves HBM3 operation of up to 8.4 Gbps leveraging over 30 years of high-speed signalling expertise, and a strong history of 2.5D memory system architecture design and enablement. In addition to the fully integrated HBM3-ready memory subsystem, Rambus provides its customers with interposer and package reference designs to speed their products to market.

“With the performance achieved by our HBM3-ready memory subsystem, designers can deliver the bandwidth needed by the most demanding designs,” said Matt Jones, general manager of Interface IP at Rambus. “Our fully-integrated PHY and digital controller solution builds on our broad installed base of HBM2 customer deployments and is backed by a full suite of support services to ensure first-time-right implementations for mission-critical AI/ML designs.”

Let’s look at some benefits of the Rambus HBM3-ready Memory Interface Subsystem:

  • Supports up to 8.4 Gbps data rate delivering bandwidth of 1.075 Terabytes per second (TB/s)
  • Reduces ASIC design complexity and speeds time to market with fully integrated PHY and digital controller
  • Delivers full bandwidth performance across all data traffic scenarios
  • Supports HBM3 RAS features
  • Includes built-in hardware-level performance activity monitor
  • Provides access to Rambus system and SI/PI experts helping ASIC designers to ensure maximum signal and power integrity for devices and systems
  • Includes 2.5D package and interposer reference design as part of IP license
  • Features LabStation development environment that enables quick system bring-up, characterization, and debugs
  • Enables the highest performance in applications including state-of-the-art AI/ML training and high-performance computing (HPC) systems

Many in the industry expect the first generation of HBM3 memory to be very similar to HBM2E, meaning that in terms of capacity it should be made up of 16Gb DRAM Dies for a total of 16 GB (8-hi stack).

However, there are reports that the memory densities of HBM3 will increase once the specifications are finalized by JEDEC. It also reported that products such as the AMD’s Instinct Accelerators based on next-gen CDNA architecture, NVIDIA’s Hopper GPUs, and Intel’s future HPC accelerators based on next-gen Xe-HPC architecture, all possess the HDM3 memory stack.

source

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.

More like this

Nvidia’s Hopper H100 appeared online in its SXM5 form...

At GTC 2022, Nvidia unveiled its Hopper architecture, revealing the H100 server accelerator but only displaying renders...

NVIDIA Hopper H100 GPU images appear online flaunting its...

NVIDIA's Hopper H100 GPU, a computational powerhouse geared for the next generation of data centers, was announced...

AMD could bring 3D Stacking with up to 8...

Moore's Law is Dead has detailed AMD Instinct MI300 GPUs, which will be powered by the next-generation...
SK Hynix announces the production of its next-generation HBM3 memory with 819 GB/s bandwidth

SK Hynix announces the production of its next-generation HBM3...

In the latest development, the south Korea-based SK Hynix has now become the first-ever in the industry...
SK Hynix’s HBM3 memory capable of reaching up to 665 GB/s of speed

SK Hynix’s HBM3 memory capable of reaching up to...

SK Hynix is best known for manufacturer cutting-edge memory modules and as such the company recently announced...

LATEST NEWS

BGMI 2025: Tournaments, Game Updates and Esports Highlights

BGMI (Battlegrounds Mobile India) is all set to get busy in 2025 with tournaments, game updates, and an intense fight for domination with top...

Netflix iOS App Now Lets Users Download Entire TV Seasons in One Tap

The streaming service Netflix has rolled out a one-tap download feature for a whole TV season in its iPhone and iPad apps. Now more...

The Greatest Rivalry India Vs Pakistan Trailer: Netflix’s Upcoming Cricket Documentary Set to Thrill Fans

The Greatest Rivalry India Vs Pakistan Trailer Here: Netflix has finally dropped the much-awaited trailer for The Greatest Rivalry: India vs Pakistan, a documentary...

 Pushpa 2: The Rule Now Streaming on Netflix in Kannada

Pushpa 2: The Rule Now Streaming on Netflix The highly anticipated sequel to Allu Arjun's blockbuster, "Pushpa 2: The Rule," has finally made its digital...

Featured