Overview
This LPDDR4/4X/5 PHY is a memory-side interface IP normally found implemented within commodity DRAM products. Green Mountain Semiconductor's LPDDR4X/LPDDR5 combo IP provides the unique opportunity to transmit data between a variety of devices such as AI coprocessors, in-memory compute solutions and emerging memory products.
This is a memory side (Slave-side) interface for AI processors and other ASICS seeking the latest high speed, low power LPDDR interface protocols for general purpose data transfer, while adhering to the well known and well defined LPDDR4X and LPDDR5 standard as specified by JEDEC.
This IP is designed for 7nm TSMC but can be ported to other logic processes. It is also suitable for a wide variety of memories such as DRAM, SRAM as well as emerging memories including non-volatile memories, with appropriate modifications.
Learn more about LPDDR IP core
JEDEC has just released the LPDDR6 specification that is expected to take the LPDDR DRAM market to new heights with data transfer speeds that can reach up to 14.4Gbps, which is a 50% improvement over LPDDR5X speeds. LPDDR6 device density can range from 4 GB to 64 GB.
This solution brief highlights how Samsung’s mainstream 8nm process node and LPDDR5X (LP5X) memory together deliver an optimized balance of performance, bandwidth, power efficiency, and cost. The combination enables scalable AI inference across diverse computing environments, from edge devices and AI appliances to emerging inference-focused AI infrastructure, making advanced AI deployment more accessible for startups, fabless innovators, and OEMs alike.
Next-generation automotive systems are advancing beyond the limits of currently available technologies. The addition of advanced driver assistance systems (ADAS) and other advanced features requires greater processing power and increased connectivity throughout the vehicle.
Understand memory-aware ASIC design for AI workloads, from memory hierarchy and dataflow to NoC, DMA, and hardware-software co-design.
As AI workloads continue to diversify, the systems that support them are evolving just as quickly. AI is no longer confined to the hyperscale data center. It is moving to the factory floor, into vehicles, and increasingly to the edge, where power, cost, and form factor constraints can matter just as much as raw performance.
