Options

When deploying GPUs in a high-performance computing (HPC) environment, customers face substantial obstacles and inefficiencies caused by the need to port and refactor code. Their efforts are further hampered by proprietary GPU programming environments that prohibit portability between GPU vendors and often result in inconsistency between CPU and GPU implementations. The need for GPU-level memory bandwidth, at scale, and sharing code investments between CPUs and GPUs for running a majority of the workloads in a highly parallelized environment has become essential. 

Intel Data Center GPU Max Series is designed for breakthrough performance in data-intensive
computing models used in AI and HPC.

Are you looking for a simple way to migrate to 10G high performance data center with low cost and low power in your data center? The HPE Ethernet 10Gb 2-port X550-AT2 Adapter deliver high performance and low latency for Ethernet traffic to and from the server with minimal CPU resources required. Processing of I/O packets on the adapter frees up CPU to do other more important tasks. For environments such as Telco/NFV, high-performance computing (HPC) and ecommerce that deal with large volumes of small packet traffic, these adapters accelerate small packet processing by bypassing processing in the host O/S kernel.

Do you want to improve your performance and bandwidth without increasing your hardware complexity?

The HPE Synergy 6820C 25/50Gb Ethernet Adapter provides more than two times the bandwidth compared to 10/20Gb adapters. Therefore, it simplifies hardware implementation by reducing the number of adapters, associated switches and compute modules required to achieve higher bandwidth.

The networking acceleration on this adapter alleviates the CPU utilization, increases the number of virtual machines (VMs) placed on each server, and improves the cloud scale efficiency. Pairing with the HPE Synergy high-density 50G Ethernet switching solution, this adapter provides high throughput and low latency to implement intelligent and flexible networks.

Do you require SDN features, low-latency and high-bandwidth in your data center? TheHPE Ethernet 10Gb 2-port FLR-SFP+ X710-DA2 Adapter offers the latest SDN features in a FlexibleLOM form factor. The HPE Ethernet 10Gb 2-port FLR-SFP+ X710-DA2 Adapter is a feature rich, low power dual port PCIe v3.0 x8 lane adapter designed for use in HPE ProLiant Rack and Density Optimized servers.

Do you need an adapter that is big on capability and small on CPU utilization?

The HPE Ethernet 10Gb 2-port 537SFP+ FLR Adapter has been designed to efficiently deliver high performance, low latency an minimal CPU utilization by offloading many I/O processing within the adapter. Low latency dramatically improves application performance and frees up CPU resources in the host allowing it to do tasks that are more important.

Do you have demanding workloads that require low latency? The HPE Ethernet 100Gb 1-port QSFP28 MCX515A-CCAT Adapter supports 100 Gb bandwidth that allows more network traffic and thus is ideal for telcos, cloud and service providers. This adapter has security features to keep out rogue updates that could slow you down. Plan for the future and get in the express lane with the 100 Gb adapter.

Do you need low power, high port density 1GbE network adapters for virtualization and server consolidation?

The HPE Ethernet 1Gb 4-port FLR-T BCM5719 Adapter for select HPE ProLiant servers features the next generation of Ethernet integration, offering a single-chip solution in a FlexibleLOM form factor, which offers the option for future upgrades to 10GbE networking speeds. It delivers full line-rate performance across all ports with low power consumption. The adapter is ideal for virtualization, security, server consolidation, network segmentation, and other applications requiring maximum throughput and port density. The total aggregate throughput of 4 Gb meets the needs of customers desiring high bandwidth. The adapter supports advanced features such as Large Send and Receive offload capability, TCP checksum and segmentation, VLAN tagging, MSI-X, jumbo frames, IEEE 1588, HPE Sea of Sensors 3D and virtualization features such as VMware NetQueue and Microsoft VMQ.

Do you want flexibility while being efficient? The HPE Synergy 4820C 10/20/25Gb Converged Network Adapter provides a path for 10G today and up to 25Gb tomorrow for Financial, trading, Telco and cloud service providers. This adapter converges Ethernet and FCoE onto a single connection simplifying hardware and reducing costs. Pairing with the Hewlett Packard Enterprise Synergy Switch solution, this adapter provides high throughput and low latency to implement intelligent and flexible networks connecting the server to the switch.

Need a low cost, low power 1GbE adapter for your HPE ProLiant server?

The HPE Ethernet 1Gb 4-port BASE-T I350-T4V2 Adapter features a single chip solution, further reducing power requirements for 4 ports of 1 Gb Ethernet. The HPE Ethernet 1Gb 4-port BASE-T I350-T4V2 Adapter is PCI Express 2.1 x4 based on the latest Intel Ethernet Controller I350 designed for use with select HPE ProLiant rack, tower and density optimized servers.

Would you like to increase your network’s I/O performance while simplifying the connection between HPE Synergy Compute modules and complex storage environments? Designed for the HPE Synergy Composable Fabric, HPE Synergy 5830C 32Gb Fibre Channel (FC) Host Bus Adapter (HBA) connects Synergy Compute resource pools to SANs over 32 Gb native FC fabrics. This provides high-performance due to low latency connectivity to the HPE Synergy Virtual Connect FC modules and Brocade® FC Switch modules. The HPE Synergy 5830C supports advanced virtualization, security, dual-port isolation, dynamic power management and low CPU utilization features. Provisioning and updating HPC Synergy 5830C are easy, reliable and fast using HPE Synergy software-defined, template-driven, frictionless operation. The HPE Synergy Composer automatically coordinates proper firmware and driver including staging updates to execute and avoid workload disruption.