|
contents | hardware | |||||||
| PLX Launches PCI Express Switches with Integrated DMA PLX Technology announces three new PCIe switch devices with a innovative architecture that features an integrated direct memory access (DMA) engine. Each switch provides four DMA channels to support the high data rates required in storage systems, servers, networking, control plane and embedded markets. By offloading the DMA function typically required of the processor, PLX's DMA-capable switches increase system performance and create a wide range of new options for next-generation PCIe designs. The ExpressLane PEX 8619 (16 lanes, 16 ports), PEX 8615 (12 lanes, 12 ports) and PEX 8609 (eight lanes, eight ports) PCIe Gen 2-compliant switches offer flexible and configurable ports ranging from x1 to x4 on all three devices, with up to x8 on the PEX 8619. The four DMA channels can support high-speed data transfers between I/O devices connected to any of the available ports, while maintaining independence from the unique transparent switch functionality with up to 3Gbps throughput per DMA channel. Additional features include a low latency maximum of 140ns and power requirements down to 1.2W typical (8609), with quality of service (QoS) by means of two virtual channels (VCs) per port, spread spectrum clock (SSC) isolation via dual clock domains, and end-to-end guaranteed data integrity. The DMA engine in these devices implement a descriptor ring approach, while each of the four DMA channels can saturate a x8 link at Gen 2 speeds (up to 4GB/s) in one direction. Each descriptor provides support for large transfer sizes (up to 128MB) giving the user the ability to perform very large data transfers in any direction (memory to device, device to device, memory to memory). Descriptors can exist in host memory or, alternatively, inside the DMA switch. Up to 256 descriptors are supported internally in PLX DMA switches, which also support 32-bit and 64-bit transfers as well as programmable QoS. The PEX 8619, PEX 8615 and PEX 8609 will be sampling in Q4 2008, with full production delivery slated for Q1 2009. write your comments about the article :: © 2008 Computing News :: home page |