Name: Intel OPA 100 Series HFA - Network adapter - PCIe 3.0 x16 1 - for System x3550 M5; x3650 M5; ThinkSystem SD530; SD630 V2; SR530; 00WE027
SKU: 449771

Product Details ＋

Intel OPA 100 Series HFA - Network adapter - PCIe 3.0 x16 1 - for System x3550 M5; x3650 M5; ThinkSystem SD530; SD630 V2; SR530; SR550; SR630; SR650; SR670

Advanced data transfer and Quality of Service (QoS)
With OPA, data is segmented into 65-bit Flow Control Digits (FLITs) which are assembled into larger Link Transfer Packets (LTPs) for efficient wire transfer. By managing traffic at the FLIT level, OPA edge and director switches can more effectively optimize latency, throughput and resiliency for all types of traffic.
Scalable performance
The on-load architecture provides investment protection by allowing fabric performance to scale along with the increased performance of Intel Xeon processors and Intel Xeon Phi coprocessor adapters.
Open Fabrics Alliance software
The OPA adapters are compatible with software from the Open Fabrics Alliance (OFA) which means that many existing HPC applications will be able to operate over an OPA fabric.
Enhanced reliability
The advanced design in these adapter means there is no need for data path firmware or external memory, which enhances reliability. The Intel OPA adapters have ECC protection on all internal SRAMs and parity checking on all internal buses. Equally important, the stateless design is inherently more resilient to adapter and fabric failures as it minimizes its reliance on the connection state. Passive heat sinks eliminate the risk of fan failures.

Intel Omni-Path Architecture (OPA) is a family of PCIe adapters, switches, cables, and management software that provides high-performance connectivity between servers. It is designed for high performance computing (HPC) workloads and has the ability to scale to thousands of nodes.
The PCIe adapters, known as Omni-Path Host Fabric Adapters (HFAs) use an advanced connectionless "on-load" design that uses the processing capacity of Intel Xeon processors in the server to automatically scale fabric performance as server processor core counts increase.