Nvidia

GigaIO FabreX now natively supported in NVIDIA Bright Cluster Manager 9.2

GigaIO FabreX now natively supported in NVIDIA Bright Cluster Manager 9.2

SAN DIEGO, May 19, 2022 – GigaIO, provider of an open, rack-scale computing platform for large-scale workflows, today announced that GigaIO FabreX for Composable Infrastructure is now supported loads natively in NVIDIA Bright Cluster Manager 9.2. The integration, led by NVIDIA in conjunction with GigaIO, ensures customers can create easy-to-manage, platform-independent compute clusters that scale in minutes to handle the most demanding AI and HPC workloads. demanding.

This new integration is an example of GigaIO’s strategy to provide an open platform that allows customers to access the benefits of composable infrastructure through the enterprise-class tools they already use. “Through our native integration strategy into leading software tools such as NVIDIA Bright Cluster Manager, our goal is to be invisible to data center managers so their users can seamlessly submit jobs and don’t have you don’t even need to know the magic of our software-defined software-on-the-fly hardware reconfiguration resources,” said Alan Benjamin, CEO of GigaIO.

“Companies building AI and HPC computing infrastructures are looking for solutions that deliver performance, flexibility and efficiency,” said Charlie Boyle, vice president of DGX Systems at NVIDIA. “With native support for Bright Cluster Manager 9.2, GigaIO FabreX customers can now compose and manage their compute systems to meet the needs of unique workloads from a single management interface.”

GigaIO’s universal dynamic memory fabric, FabreX, allows an entire server rack to be treated as a single compute resource. Resources normally located inside a server, including accelerators, storage, and even memory, can now be grouped together in accelerators or storage enclosures, where they are available to all servers in a rack . These resources continue to communicate through a native PCIe memory structure, as they would if they were connected to the server’s motherboard.

NVIDIA Bright Cluster Manager is an enterprise-class software solution that simplifies the creation and management of HPC clusters from edge to core to cloud, transparently to the customer, by combining provisioning, monitoring and management capabilities in one tool. Version 9.2 extends the goals of complexity elimination and flexibility by adding built-in support for composable infrastructure using GigaIO FabreX, where nodes can now be composed using Bright Cluster Manager Shell or BrightView.

Autoscaling in Bright Cluster Manager creates a dynamic and versatile infrastructure, and FabreX extends this agility to every piece of hardware in a rack by enabling the creation of composable GigaIO GigaPods and GigaClusters with cascaded and interconnected switches. PCIe devices can be monitored and verified, as well as grouped and assigned to nodes within a cluster using Bright Cluster Manager. Clusters can have several different structures defined, and Bright Cluster Manager streamlines the process of setting up the structure.

FabreX support with NVIDIA Bright Cluster Manager allows users to manage more workloads while maximizing resource utilization, minimizing costs, and managing everything from a single, state-of-the-art user interface. Cloud-like agility is now easier for on-premises infrastructure, enabling cloud bursting as needed within a single interface.

GigaIO’s dedication to native integration with software tools such as Bright Cluster Manager provides a best-in-class experience for customers who can continue to use their favorite tools without having to change their software stack. Native integration eliminates the need for customers to rely on plug-ins with limited capabilities, learn new software, manage yet another window, and pay additional license fees per node.

This development reinforces GigaIO’s close collaboration with NVIDIA. GigaIO is a member of the NVIDIA Partner Network (NPN) since 2020 and was accepted last month in Creation of NVIDIA, a program designed to encourage cutting-edge startups. GigaIO shares many common customers with NVIDIA, including the Texas Advanced Computing Center and the San Diego Supercomputer Center.

Dr. Frank Würthwein, director of the San Diego Supercomputer Center, is one of the beneficiaries of this collaboration. “Our research demands that we aggregate disparate computational elements, such as GPUs, x86 processors, and storage systems into highly usable and reconfigurable systems,” he said. “GigaIO’s FabreX technology combined with NVIDIA Bright Cluster Manager allows these elements to be dynamically brought together in a high-performance, ultra-low latency interconnect while enabling the coexistence of separate, interference-free workflows on the same infrastructure. »

Learn more about native FabreX integration for composable infrastructure in NVIDIA Bright Cluster Manager 9.2, available now.

About GigaIO

Based in Carlsbad, CA, GigaIO provides the world’s only open rack-scale computing platform, delivering the elasticity of the cloud at a fraction of the TCO (total cost of ownership). With its universal dynamic memory fabric, FabreX, and innovative open architecture using standard PCI Express/soon to be CXL technology, GigaIO breaks down the constraints of the server enclosure, freeing up resources to shorten time to results. Contact [email protected]to visit www.gigaio.com.


Source: GigaIO