HBM Controller IP

Delivering ultra high-bandwidth, low-latency memory performance

HBM Memory Controller IP

Rambus High-Bandwidth Memory (HBM) 3E/3 and 2E/2 controller IP provide high-bandwidth, low-latency memory performance for AI/ML, graphics and HPC applications. 

Version Maximum Data Rate (Gb/s) Controller
HBM3E / HBM3 9.6 Download HBM3 Product Brief Product Brief
HBM2E / HBM2 3.6/2.0 Download HBM3 Product Brief Product Brief

HBM3E / HBM3 & HBM2E Controller IP

Product Brief
Product Brief
Speed Bins (Gb/s) Up to 9.6/8.4Up to 3.6/2.0
Channel Densities (Gb)Up to 32Up to 24
Channels | Pseudo-Channels16 | 328 | 16
DRAM StacksUp to 16Up to 12
PHY InterfaceDFI StyleDFI Style
PHY Independent ModeYesYes
Refresh Management SupportYes 
Look-Ahead Command Processing
for Minimum Latency
Integrated Reorder FunctionalityYesAdd-on core
Self-refresh and Power-down
Low Power Modes
RAS FeaturesYesYes
Built-in Activity MonitorYesYes
DFI CompatibleYesYes
End-to-end Data ParityYesYes
Interface to LogicNative or AXINative or AXI

HBM3E Memory: Break Through to Greater Bandwidth

HBM3E Memory: Break Through to Greater Bandwidth

Delivering unrivaled memory bandwidth in a compact, high-capacity footprint, has made HBM the memory of choice for AI training. HBM3 is the third major generation of the HBM standard, with HBM3E offering an extended data rate and the same feature set. The Rambus HBM3E/3 Controller provides industry-leading performance to 9.6 Gb/s, enabling a memory throughput of over 1.23 TB/s for training recommender systems, generative AI and other compute-intensive AI workloads.

HBM3E Memory Subsystem

HBM is a high-performance memory standard that features reduced power consumption and a small form factor. It combines a 2.5D/3D architecture with a wider interface at a lower clock speed (as compared to GDDR6) to deliver higher overall throughput at a higher bandwidth-per-watt efficiency for AI/ML and high-performance computing (HPC) applications. 

HBM3E Memory Subsystem Example
HBM3E Memory Subsystem Example

Rambus HBM memory controllers support data rates up to 9.6 Gb/s per data pin (HBM3E). With its 1024-bit interface and this maximum data rate, it can deliver memory throughput of 1.23 TB/s.

Rambus HBM controllers can be paired with 3rd-party or customer PHY solutions.

Ask the Experts: HBM3E Memory Interface IP

In this episode of Ask the Experts, we discuss HBM3E memory with Nidish Kamath, director of product management for memory interface IP at Rambus, and Frank Ferro, group director of memory and storage IP at Cadence. Learn about the evolution of the HBM in the market, challenges of meeting higher performance targets of HBM3E, and IP solutions you can integrate into your project today.

Frequently Asked Questions about HBM

HBM is a high-performance memory standard that delivers extremely high bandwidth at excellent power efficiency by employing a wide 1024-bit interface at relatively low data rates.

HBM3 memory raises the data rate to 6.4 Gb/s to deliver 819.2 GB/s of bandwidth per HBM3 memory device. The Rambus HBM3E digital controller can operate at 9.6 Gb/s to provide additional design margin.

HBM memory offers excellent bandwidth and capacity with superior power efficiency that makes it an ideal solution for AI/ML training workloads.

To achieve its high bandwidth, HBM uses a 1024-bit wide data interface. This is far more than can be supported on a standard PCB used by traditional (2D) memory solutions. So, HBM employs a silicon interposer where traces can be very finely etched. This structure is called 2.5D. In addition, HBM uses 3D-stacked memory devices, giving HBM a 2.5D/3D architecture.

HBM3 memory is ideal for data-intensive applications like AI/ML, graphics, and HPC, where large amounts of data need to be processed at very high bandwidth.

HBM3 and GDDR6: Memory Solutions for AI

HBM3E and GDDR6: Memory Solutions for AI

AI/ML changes everything, impacting every industry and touching the lives of everyone. With AI training sets growing at a pace of 10X per year, memory bandwidth is a critical area of focus as we move into the next era of computing and enable this continued growth. AI training and inference have unique feature requirements that can be served by tailored memory solutions. Learn how HBM3E and GDDR6 provide the high performance demanded by the next wave of AI applications.

Rambus logo