The Supermicro Multi-Node BigTwin® HCI System with 4th Gen Intel® Xeon® Scalable Processors Provides 2X the I/O Bandwidth with PCIe 5.0 and up to 1.5X the memory bandwidth with DDR5 and Supports Built-in CPU Accelerators for Workload Specific Performance Gains
Supermicro, Inc. (NASDAQ: SMCI), a Total IT Solution Provider for Cloud, AI/ML, Storage, and 5G/Edge, today announced a new VMware vSAN solution optimized to run enterprise class hyperconverged virtualized workloads. As virtualized workloads become more advanced, processing power and storage performance requirements increase, requiring greater capacity to meet application SLAs and maximize virtual machine density. This solution also utilizes the latest Intel AMX accelerator for AI workloads.
When compared to the Supermicro X11 BigTwin, benchmark testing conducted by Supermicro showed up to 4.7X higher IO throughput and 8.2X lower latency on the HCIBench benchmark, up to 4.9X faster image classification inference throughput on the ResNet50 model and up to 4X faster natural language processing throughput on BERT-Large model. In addition, the superior power and efficiency of the Supermicro X13 BigTwin architecture can deliver up to 3X cost and performance improvement within the same node footprint compared to a similar deployment based on older generation Supermicro systems, creating a compelling case for organizations to upgrade their aging infrastructure.
“Supermicro continues to lead the HCI industry with the first Gen5 vSAN solution based on the 4th Gen Intel Xeon Scalable processors,” said Charles Liang, president and CEO, Supermicro. “Our industry leading Supermicro BigTwin platform supports the latest generation of flash storage media in dense, multi-node form factors optimized for HCI to run vSAN out of the box. Supermicro’s X13 BigTwin and vSAN customers can benefit from 4.7X higher throughput and 8.2X lower latency on the HCIBench benchmark with gen-over-gen performance improvements. This solution enables customers to run more instances on the same hardware and rack space, improving performance and utilization while reducing costs.”
Supermicro, Inc. (NASDAQ: SMCI), a Total IT Solution Provider for Cloud, AI/ML, Storage, and 5G/Edge, today announced a new VMware vSAN solution optimized to run enterprise class hyperconverged virtualized workloads. As virtualized workloads become more advanced, processing power and storage performance requirements increase, requiring greater capacity to meet application SLAs and maximize virtual machine density. This solution also utilizes the latest Intel AMX accelerator for AI workloads.
When compared to the Supermicro X11 BigTwin, benchmark testing conducted by Supermicro showed up to 4.7X higher IO throughput and 8.2X lower latency on the HCIBench benchmark, up to 4.9X faster image classification inference throughput on the ResNet50 model and up to 4X faster natural language processing throughput on BERT-Large model. In addition, the superior power and efficiency of the Supermicro X13 BigTwin architecture can deliver up to 3X cost and performance improvement within the same node footprint compared to a similar deployment based on older generation Supermicro systems, creating a compelling case for organizations to upgrade their aging infrastructure.
“Supermicro continues to lead the HCI industry with the first Gen5 vSAN solution based on the 4th Gen Intel Xeon Scalable processors,” said Charles Liang, president and CEO, Supermicro. “Our industry leading Supermicro BigTwin platform supports the latest generation of flash storage media in dense, multi-node form factors optimized for HCI to run vSAN out of the box. Supermicro’s X13 BigTwin and vSAN customers can benefit from 4.7X higher throughput and 8.2X lower latency on the HCIBench benchmark with gen-over-gen performance improvements. This solution enables customers to run more instances on the same hardware and rack space, improving performance and utilization while reducing costs.”
The Supermicro X13 BigTwin solution is available as a 2U 2-node or 2U 4-node configuration with optional liquid cooling. Each node is powered by up to two 4th Gen Intel Xeon Scalable processors and up to 4TB of DDR5-4800MHz memory. The system runs the latest VMWare vSAN 8.0 software stack featuring the new Express Storage Architecture (ESA) to create a single storage pool across the up to 48 NVMe drives accessible by the four nodes.
SHARE
