Super Micro Computer, Inc., a prominent provider of enterprise computing, storage, and networking solutions, revealed a comprehensive range of high-performance computing (HPC) systems at the SuperComputing 2020 (SC20) event.
Held virtually from November 17 to 19, the showcase included product demonstrations and speaking sessions led by Supermicro's senior representatives focusing on recent HPC deployments.
Innovative Supermicro HPC Systems
Supermicro has developed a diverse selection of advanced HPC systems, custom-designed to cater to specific software, storage, and cooling/power requirements. This adaptability facilitates the rapid implementation of field-proven HPC solutions, optimised for various industries and applications.
Supermicro has developed a diverse selection of advanced HPC systems
In a statement, Charles Liang, president and CEO of Supermicro, highlighted the company's focus on continually advancing HPC technologies to support computationally demanding scientific research and discovery. He noted, "Supermicro is actively innovating to provide customer solutions for these dynamic markets." He also emphasized their commitment to resource-saving architecture aiming to reduce power consumption and total cost of ownership as well as IT waste and environmental impact.
Supermicro Event Sessions at SC20
One of the key sessions scheduled included "Advanced High-performance Computing Systems to Find Cure for COVID-19," outlining the capabilities of the Corona and Mammoth Clusters at Lawrence Livermore National Laboratory, which took place on Tuesday, November 17, from 2:30 p.m. to 2:45 p.m.
Another notable session, "Deploying Liquid-cooling Supercomputer Cluster at Scale in a Challenging Environment," was held on Wednesday, November 18, from 12:45 p.m. to 1:00 p.m., featuring the Ruby Cluster at the same laboratory.
Advanced Compute Solutions
Supermicro's portfolio includes advanced compute solutions like the SuperBlade® systems
Supermicro's portfolio includes advanced compute solutions like the SuperBlade® systems, offering configurations such as one-socket, two-socket, and four-socket blade servers with support for processors up to 280 watts, NVMe, 200G HDR InfiniBand switches, and Ethernet switches. These servers are suitable for enterprise, cloud, and HPC applications due to their high density, offering significant performance capabilities, including up to 40 CPUs and 40 GPUs in 8U configurations.
Moreover, Supermicro is renowned for providing an extensive selection of GPU servers tailored for artificial intelligence (AI), deep learning, and HPC workloads. With a variety of systems from 1U to 10U, capable of housing up to 40 GPUs, Supermicro’s servers deliver a substantial performance boost across a multitude of applications.
Multi-node Twin Family and Liquid-cooled Solutions
Supermicro's multi-node Twin family, a major component of their offerings, is engineered to fulfill the demands of high-density environments and shared power requisites in HPC settings. The BigTwin™ series, part of this family, is equipped to support up to 280W processors and provides six NVMe drives per node within its 2U structure, making it suitable for cutting-edge supercomputing clusters.
As HPC applications evolve and grow increasingly complex, addressing power and cooling challenges becomes ever more critical. Supermicro highlights its expanded line of liquid-cooled server solutions, designed to enhance performance for data centres focusing on computationally intensive workloads.
Super Micro Computer, Inc., a global provider of enterprise computing, storage, networking solutions and green computing, brings the industry’s broadest selection of high-performance computing (HPC) systems to SuperComputing 2020 (SC20), in Supermicro’s virtual booth, from November 17-19.
In addition to product demonstrations, Supermicro is also hosting several speaking sessions on recent HPC deployments from senior Supermicro representatives.
Supermicro offers an extensive range of leading-edge HPC systems optimised for specific software, storage and power/cooling requirements to accelerate the deployment of complete field-proven HPC solutions.
Authority comment
“Supermicro offers the most advanced HPC systems targeting computationally intensive applications for scientific research and discovery across a range of disciplines and Supermicro is actively innovating to provide customer solutions for these dynamic markets,” said Charles Liang, president and CEO of Supermicro.
“Our extensive portfolio supports our focus on Resource-Saving Architecture™ designed to help customers reduce power consumption and TCO, plus the ability to replace critical components and reuse many of the subassemblies for multiple refresh cycles to reduce IT waste and minimise TCE (Total Cost to the Environment).”
Supermicro Sessions at SC20
- Advanced High-performance Computing Systems to Find Cure for COVID-19 – The Corona and Mammoth Clusters at Lawrence Livermore National Lab (LLNL)
- Tuesday, November 17, 2020, 2:30 p.m. – 2:45 p.m.
- Deploying Liquid-cooling supercomputer cluster at Scale in a Challenging Environment – The Ruby Cluster at Lawrence Livermore National Lab (LLNL)
- Wednesday, November, 18, 2020, 12:45 p.m.-1:00 p.m.
Advanced compute solutions
Supermicro provides advanced compute solutions including the SuperBlade® systems
Supermicro provides advanced compute solutions including the SuperBlade® systems with one-socket, two-socket and four-socket blade servers supporting up to 280-watt processors, NVMe, 200G HDR InfiniBand switches, 25G/10G Ethernet switches, single or double-width GPUs, redundant AC/DC power supplies and Battery Backup (BBP®), making them ideal for enterprise, cloud and HPC applications.
Critical to HPC is the industry's foremost density that SuperBlade provides (10 x 4-socket or 20 x 1 or 2-socket blade servers), delivering up to 40 CPUs and 40 GPUs in 8U.
GPU servers optimised for AI
Supermicro offers the industry’s broadest selection of GPU servers optimised for AI, Deep Learning and HPC workloads, including a full line of systems from 1U up to 10U that support up to 40 GPUs with two single-width PCI-E GPUs in one node for SuperBlade in a single system.
Supermicro offers high-density 2U and 4U servers with interconnected 4-GPU and 8-GPU boards as well as 8 PCI-E based on the four- and eight-way boards.
With several GPUs, also available in PCIe form factor, customers can also expect a major performance boost across Supermicro’s extensive portfolio of multi-GPU servers including 1U Ultra and BigTwin™ and embedded systems supporting GPUs; as well as GPU blade modules up to 2 GPUs per module for our 8U SuperBlade®.
Supermicro offers the industry’s widest and deepest selection of GPU systems to power supercomputing applications.
Supermicro’s multi-node Twin family
Supermicro’s multi-node Twin family was designed to meet the density and shared power requirements of HPC environments
Supermicro’s multi-node Twin family was designed to meet the density and shared power requirements of HPC environments. Supermicro’s new BigTwin™ multi-node family delivers density, shared power and full performance for HPC.
The BigTwin is a 2U four dual-processor node system with six NVMe drives per node. This system is capable of supporting up to 280W processors. Also, the H12 BigTwin family of servers is optimised to deliver a new level of integration and superior performance for modern supercomputing clusters.
Liquid-cooled server solutions
HPC applications are continuing to grow in complexity as they unlock new scientific insights. HPC deployments, especially data-intensive projects, can yield sophisticated environmental challenges (power and cooling).
Supermicro is featuring its expanded offering of liquid-cooled server solutions to address the power and cooling challenges for computationally intensive workloads in today’s HPC data centres.
