Genuine Cornelis Networks through authorized distribution — TAA country-of-origin verified and NDAA §889 screened before checkout.
SKU
100HFA016LSN
Category
Host Bus Adapters
Overview
Cornelis Networks provides the industry's leading accelerated host fabric adapters.
Omni-Path Express adapters cost-effectively deliver high bandwidth and use advanced technologies to meet the key challenges to application performance, maximizing cluster scalability and message rate while minimizing average and tail latency.
Cornelis Omni-Path Express scale-out interconnect
Unprecedented demands on the scale-out interconnect are being driven by advances in artificial intelligence, high performance data analytics, and traditional modeling and simulation environments, coupled with extremely capable processing and storage infrastructures.
Cornelis Omni-Path Express is the next generation of high performance fabrics, a proven hardware foundation combined with the OpenFabrics Interfaces (OFI) framework, that delivers the industry's lowest latency, highest message rate, and best collectives performance, all at the industry's lowest CPU utilization.
Accelerated application performance at scale
Cornelis Omni-Path Express Accelerated Host Fabric Adapters provide a perfect semantic match between the requirements of real-world applications and the scale-out fabric, maximizing scalability and performance at the industry's lowest price point.
These adapters ensure optimal application performance by delivering key features for efficiency, including dispersive routing and congestion control. These features are complemented by a unique sub-link layer architecture that enables Packet Integrity Protection (zero latency protection against bit transmission errors).
These features, together with advanced Virtual Fabrics support, provide the unique interconnect capabilities to deliver industry-leading application performance and manageability at scale.
Highlights
Accelerated application performance at scale
Industry leading price-performance
Advanced sub-link layer capability eliminating link protection and tail latency penalties
100 Gbps in standard format
PCIe x16 host interface
Over 160M MPI msg/sec
Sub-microsecond MPI latency Advanced features
Dispersive Routing
Packet Integrity Protection
Congestion Control
Virtual Fabrics Support
Lowest end-to-end latency at scale
Best collectives performance
Balanced functionality between CPU and network with low CPU utilization