GigaIO Announces Next Generation Edge-to-Core Product Line for Scale-Up AI and HPC Infrastructure

GigaIO’s multi-interconnect, multi-accelerator fabric technology now expanded to PCIe Gen 5.

Carlsbad, California, November 14, 2024 – GigaIO, an award-winning provider of open workload-defined infrastructure for AI and accelerated computing, today announced its next generation product lineup, which meets the demands of a multi-accelerator, multi-interconnect AI and HPC world. GigaIO’s core technology FabreX^TM, an AI memory fabric, now incorporates PCIe Gen 5 to support state-of-the-art performance. FabreX powers GigaIO’s fully engineered solutions, from SuperNODE^TM, the only 32-GPU server, to Gryf^TM, the first suitcase-sized AI supercomputer.

“We are proud of the team for solving previously intractable problems,” said Alan Benjamin, CEO of GigaIO. “With artificial intelligence requiring ever more domain-specific hardware, and interconnects being recognized as both a key bottleneck and an enabler of next-gen infrastructure, GigaIO offers the most complete multi-accelerator, multi-interconnect fabric solutions from the data center to the very edge.”

GigaIO’s new product line-up includes:

• Accelerator Pooling Appliances (APAs) for PCIe-format accelerators, in both 8- and 16-slot versions. APA choice provides flexibility and future-proofing to organizations that anticipate the need to integrate yet-to-be launched specialized cards, such as those from d-Matrix.
• APAs for both NVIDIA SXM-format and AMD OAM-format 8-GPU chassis. With SuperNODE, several of these APAs can now be combined to function as if they were part of a unified, rack-scale server. This breakthrough configuration dramatically improves data throughput and latency, providing performance levels previously unattainable on competing multi-node solutions.
• A PCIe Gen 5 24-port switch.
• Adapter cards for both APAs and host servers.
• PCIe Gen 5 cables with the new QSFP-DD connectors, in both copper and optical fiber versions. The optical solution is being demoed at SC24 by Microchip in the PCI-SIG booth (#3354).

“At Hot Aisle, we’re thrilled to partner with GigaIO in bringing their groundbreaking multi-interconnect, multi-accelerator solutions into our Tier 5 high-end data centers,” said Jon Stevens, CEO of Hot Aisle Inc. “As customer demand for scalable, high-performance compute continues to surge, GigaIO’s FabreX technology aligns perfectly with our mission to provide flexible, future-proof infrastructure for AI and HPC. By leveraging GigaIO’s PCIe Gen 5-enabled innovations, we’re able to offer unparalleled performance, ensuring our clients have access to the best technology to meet today’s AI and data-driven challenges.”

GigaIO’s continued technology leadership is exemplified by its contributor-level memberships in UALink and the CXL Consortium, as FabreX delivers the only open, vendor-agnostic, and accelerator-agnostic scale-up solution for accelerated computing available today. As a multi-dataplane interconnect fabric based on workhorse-standard PCIe, FabreX manages CXL, Infinity Fabric, NVLink, and UALink, and expands options to use multiple brands, models, and types of accelerators, from GPUs to FPGAs and inference cards.

Early access to GigaIO’s next-generation product line will be available to select partners and customers in Q1. Stop by the GigaIO booth (#2945) at SC24 in Atlanta to see the SuperNODE at work or learn more here.

About GigaIO
GigaIO is an award-winning provider of open workload-defined infrastructure for AI and accelerated computing, delivering innovative solutions for AI and HPC challenges. GigaIO’s portfolio features two edge-to-core accelerated AI systems: SuperNODE, the world’s only 32 GPU single-node AI supercomputer, and Gryf, a highly portable and scalable AI supercomputer solution that provides meaningful computing and storage at the edge. Both utilize GigaIO’s AI memory fabric, FabreX, which seamlessly composes rack-scale resources and integrates natively into industry-standard tools. These solutions provide near-perfect scaling for LLM inference, RAG, and fine-tuning, as well as breakthrough performance in engineering and scientific workloads. Visit www.gigaio.com, or follow on Twitter (X) and LinkedIn.

Contact

Danica Yatko
760-487-8395
danica@xandmarketing.com