BittWare, a Molex company, announced the expansion of its IA-Series of FPGA accelerators featuring Intel® Agilex™ FPGAs. The BittWare IA-Series of FPGA accelerators are designed to help customers develop and deploy next-gen edge and cloud applications, with greater flexibility and speed while reducing risk.
The new IA-420F low-profile PCIe card and IA-220-U2 computational storage processor join the IA-840F to provide a choice of three form factors popular in data center and network infrastructure. The trio of products are fully PCIe Gen 4 compliant and include application reference designs, in addition, to support for the oneAPI™ programming model, which allows hardware developers the ability to create domain-specific FPGA platforms and application developers to build cross-architecture, single-source compilation designs.
“BittWare is excited to leverage Intel’s Agilex technology to support three of the most prevalent accelerator form factors used for data center and network infrastructure,” said Craig Petrie, vice president, Sales, and Marketing of BittWare. “As a part of Molex, we are uniquely positioned to drive technology innovations in enterprise-class products, which empowers our customers to deploy Intel Agilex FPGA-based solutions at the card and server-level with reduced risk and cost.”
According to Gartner’s Top 10 Strategic Technology Trends for 2020: Empowered Edge, by 2023, more than 50% of enterprise-generated data will be created and processed outside the data center or cloud, up from less than 10% in 2019. FPGAs play an increasing role at the edge where real-time data with critical latency, bandwidth, and security requirements need to be processed efficiently. Within the cloud, FPGA heterogeneous computing solutions are required to process vast amounts of both structured and unstructured data efficiently through high-performance computing and big data analytics to derive value and orchestrate smart devices in the field.
Expanded Accelerator Family Increases Deployment Flexibility
The new IA-420F low-profile PCIe card is designed to address SmartNIC and computational storage use cases. This half-height, half-length single-width PCIe card is compatible with almost all server and edge platforms. The IA-420F supports multiple 100G network ports for network and sensor processing applications with high data ingest requirements. A full 16-lanes of PCIe Gen 4 provides customers with up to twice the bandwidth normally available for FPGA-augmented systems, thereby making it compelling for data analytic workloads.
The IA-220-U2 computational storage processor is designed to address NVMe computational storage workloads. Conforming to the popular U.2 form factor, the IA-220-U2 features an Intel Agilex FPGA with an embedded ARM processor subsystem and support for PCIe Gen 4. This energy-efficient, flexible compute node can be deployed within conventional U.2 NVMe storage arrays and is chassis front-serviceable. The IA-220-U2 also can serve as a deployment-friendly computational storage processor within data center platforms that have a surplus of U.2 slots, but no available PCIe slots.
The IA-420F and IA-220-U2 join the IA-840F full-size PCIe card, which was introduced in November 2020, to provide a complete family of solutions for the most demanding data center environments.
“Intel Agilex FPGAs are delivering industry-leading performance and power efficiency, and with cross-platform tools including the oneAPI™ toolkits, offer customers the adaptability and agility to innovate for diverse workload transitions in 5G, network, cloud, and edge,” said Deepali Trehan, vice president Data Platforms Group and General Manager, Programmable Solutions Group Product Marketing at Intel. “BittWare and Molex, both members of the Intel Partner Alliance program, solve some of the toughest data-centric compute, network, and storage challenges. By taking advantage of Intel Agilex FPGAs and the oneAPI toolkits, BittWare can successfully help transition enterprise customers from proofs-of-concept to volume deployments while addressing qualification, validation, life cycle management, and support requirements.”