Super Micro Computer, Inc. (SMCI) announces the delivery of its most advanced 4 -sockets in the sector, designed for business and large -scale databases. Equipped with the latest Intel Xeon 6 processors with performance-cooks (P-Cores), these new servers are ready for the CXL 2.0 standard and are ideal for high performance calculation (HPC), critical workloads and databases in memory. Other applications will also benefit from these systems, in particular those requiring a very significant memory capacity associated with a massive calculation power, or even scientific simulation applications strongly dependent on CPU performance with low latencies.
Key component of Data Center Building Block Solutions® (DCBBS) (DCBBS) solutions of Supermicro, these new servers offer performance significantly higher than those of previous generations of 4 -sockets. The 4 Supermicro Sockets Servers are certified for SAP HANA and ORACLE LINUX. The large memory capacities of these systems ensure optimal performance for SAP and Oracle, making it possible to develop all the workloads on a single node, without the latency inherent in a rise in horizontal load via the network.
In addition, 2U 4 -sockets systems can accommodate two double width GPUs, while the 4U version can support up to six, for real -time IA training and inference in batch mode, in connection with business database applications. This unrivaled versatility and capacity guarantee the sustainability of SAP and Oracle deployments, faced with the rapid evolution of generative AI automation and ERP workflows. These new 4 -sockets servers simplify the creation of a high performance data center for business applications such as artificial intelligence (AI), databases, analytics, business intelligence, ERP, CRM, virtualization and scientific workloads.
By bringing together significant calculation, memory and storage resources within a single body of the operating system, these servers are capable of treating very voluminous workloads without the complexity linked to the implementation of parallel treatment on several nodes.