ALLMODELCOMPUTERNEWS : RAMBUS DDR3: Advancing the Main Memory Roadmap

DDR3: Advancing the Main Memory Roadmap
Demand for an enriched end-user experience and increased performance in next-generation mainstream computing applications is unremitting. Driven by multi-core computing, virtualization and processor integration trends, the industry needs a next-generation main memory solution capable of achieving data rates of up to 3200Mbps in the same, or lower power envelope as the current DDR3 1600MHz memory solution. The divergence of these two requirements, increasing performance while lowering power, presents a difficult challenge for future memory system designers.

In addition, next-generation memory solutions face potential bottlenecks in access efficiency and capacity, both of which have fallen as date rates increased. Memory module upgrades are the most common way to increase capacity in a system. The number of modules supported on a DDR3 memory channel drops at high data rates due to degraded signal integrity. This problem is so severe that by 1333MHz, most DDR3 memory channels can support only a single module. This makes a DDR3 memory system inadequate for most server, workstation and high-end PCs. Memory access granularity also suffers as data rates increase due to the disparity between the interface and core access speeds. The result is an increase to the core prefetch and a sub-optimal transfer size for future multi-core and graphics computing applications.


Rambus innovations such as Module Threading can provide the performance needed in memory systems beyond DDR3

In order to address these challenges, Rambus has announced a set of innovations focused on advancing single-ended signaling technologies to meet the memory system requirements of next-generation computing applications. Rambus' solution builds upon existing innovations and designs, such as FlexPhase™ circuitry, FlexClocking™ and Dynamic Point-to-Point technologies, and newly introduced innovations for Near Ground Signaling and Module Threading. When used in combination, these innovations, available for licensing, can enable future main memory systems to achieve double the data rates, 50% higher memory access efficiency, and 40% lower power consumption when compared to current DDR3 solutions.

Unleashing the Performance of High Performance Computing (HPC) with XDR™ Memory
Multi-core processing offers a means of breaking through the limitations of clock frequency scaling in order to deliver increased computing performance. Two or more cores in a processor allows for more performance while reducing power consumption per unit of work. This increased efficiency is achieved through the execution of multiple simultaneous tasks on separate cores for load balancing, security, virus protection, vector processing, etc. As a result, PCs, workstations, notebooks and servers can take advantage of efficient processing power to simultaneously run foreground applications and graphics processing while doing multi-stream HD media compression and decode, maintenance and anti-virus work in the background.

Realizing the benefits of multi-core operation requires increased memory bandwidth. Unless bandwidth is scaled commensurately with processor power, cores will be memory starved and performance will suffer. Furthermore, future memory bandwidth requirements are expected to increase dramatically. With three or more foreground applications running at 6.4GB/s each, graphics processing at 25.6GB/s, vector processing at 25.6GB/s, plus media decode and encode, security and anti-virus, PC bandwidth requirements can exceed 100GB/s.


The Rambus XDR™ memory architecture is well positioned to meet the bandwidth requirements of multi-core processors now and in the future. Current high-volume production XDR DRAM can provide single-device bandwidths of 16GB/s. A single XDIMM containing a total of 8 XDR devices could deliver a total memory bandwidth of 128GB/s. And the XDR architecture has a roadmap to device bandwidths of 48 GB/s and beyond with XDR2 DRAM. XDR memory offers the added benefit of increased signal integrity through differential signaling, better power efficiency and easier board design with Rambus innovations such as FlexPhase™ circuit technology. Multi-core technology is the next step in the evolution of computing. With the XDR memory architecture, designers can unleash the power of multi-core processors.