The demand for high performance computing (HPC) solutions continues to accelerate at an extraordinary pace. The global market size was valued at USD 39.1 billion in 2019 and is expected to grow at a compound annual growth rate (CAGR) of 6.5% from 2020 to 2027.
“The growing need for high-efficiency computing, continued diversification, and expansion of the IT industry, advances in virtualization, and rising preference for hybrid High Performance Computing (HPC) solutions are some of the factors expected to fuel the growth of the market. The ability of HPC systems to process large volumes of data at higher speeds is prompting government agencies, defense agencies, academic institutions, energy companies, and utilities to adopt HPC systems, which also bodes well for the growth of the HPC market.” — High Performance Computing Market Size, Share & Trends Analysis Report by Grand View Research.
As a result of this demand, we’re seeing private and public sector organizations pursue hardware solutions architected with innovative high performance and high efficiency server technology. Hardware vendors that develop and offer end-to-end green computing solutions for use in the data center – HPC, AI, cloud computing, enterprise IT, big data, IoT, and embedded markets – are primed to reap the benefits of this upswing in demand.
A Versatile Portfolio of Next-gen Servers
Supermicro is a leading innovator in HPC technology. The company’s A+ product line of AMD EPYC™ based systems include servers, storage, GPU-optimized, SuperBlade, and Multi-Node Twin Solutions to exactly match system requirements for challenging enterprise workloads. The servers incorporate single-socket and dual-socket system solutions designed to reduce the time to results and drive better business decisions. For example, the recently announced new 2U 2-node multi-GPU server is the perfect platform for video streaming, high-end cloud gaming, and countless social networking applications. It has market-leading system flexibility, and resulting cost savings will deliver uninterrupted performance. With innovative server designs that reduce the power required while leading the industry in performance, Supermicro leads the way with application-optimized servers, enabling modern enterprises to reduce costs and enhance their user’s experience. Supermicro is proud of the performance and efficiency gains for cloud, AI, and storage workloads, which yield the benefit offered by the company’s new generation of servers and storage systems.
Portfolio highlights include advances in rackmount server technology, featuring an all-new compact model sporting 2 x OCP 3.0 compliant Supermicro Advanced I/O Modules to provide a wide range of networking options and 10 NVMe drive bays enabling vastly more efficient transport of data between storage systems and servers, all in a 1U form factor and perfect for high-efficiency cloud computing environments.
Here is a short list of important characteristics for the comprehensive Supermicro line of servers:
- Up to 64 cores/128 threads per socket with AMD EPYC™ 7003 and 7002 Series Processors
- Up to 32 DIMMs of DDR4-3200MHz memory for up to 8TB per system
- Increased I/O throughput with PCI-E 4.0 and up to 128 lanes per socket
- Hot-pluggable U.2 NVMe storage for better application responsiveness
- 3-Year Limited Warranty and 24-hour technical support
Supermicro’s flagship performance multi-node architectures include the A+ BigTwin and TwinPro for hosting and storage-defined networking clusters; A+ SuperBlade for extreme cost-efficiency in high-density computing, and the A+ FatTwin with front and rear I/O for ultimate serviceability. In addition, there is the enormously popular A+ dual-processor GPU servers, which support up to 160 PCI-E 4.0 lanes and contain up to 8 direct-attached GPUs for more performant AI and Machine Learning frameworks.
The Supermicro A+ portfolio was specifically architected to satisfy the needs for a broad range of high-performance workloads:
- Performance and Density – The A+ product family includes a high level of performance per rack unit with more CPU cores, more GPUs, more NVMe drives, and more high-speed networking connections, all achieved through innovative system architectures and thermal designs, developed in close collaboration with AMD. A+ Servers with the latest 3rd Gen AMD EPYC™ and 2nd Gen AMD EPYC™ processors have achieved a multitude of world records in various categories.
- Maximum I/O and Lowest Latency – A+ servers were built for modern data center workloads, and fully utilize fast PCI-E 4.0 lanes with the latest technologies, including up to 200G networking, GPUDirect RDMA and Storage with ultra-fast NVMe and NVMe-oF capability, directly connected CPU-GPU, and GPU-GPU interconnection for the fastest I/O speed with lowest latency.
- Maximum Acceleration for AI, Deep Learning and HPC – Supermicro offers the broadest portfolio of GPU Systems, powered by dual 3rd Gen AMD EPYC™ Processors. Leveraging an advanced thermal design, including liquid cooling and custom heatsinks, Supermicro 2U and 4U GPU systems feature NVIDIA’s latest A100 or AMD Instinct™ MI100 GPUs in hyper-dense multi-GPU, multi-node systems. Supermicro’s flexible Advanced I/O Module (AIOM) form factor further enhances multi-GPU communication for data-hungry AI applications.
A+ Server Portfolio Highlights
Supermicro A+ servers are Better: broadest portfolio of versatile systems, Faster: maximum I/O and lowest latency, and Greener: reduced environmental impact and lower TCO. Here are some product details for the Supermicro A+ line.
Designed to deliver the highest performance, flexibility, scalability and serviceability to demanding IT environments, and to power mission-critical Enterprise workloads, the Ultra supports dual 3rd Gen AMD EPYC™ Processors and 32 DIMMS of DDR4-3200MHz memory for up to 8TB of capacity in 1U or 2U form factors.
Wide range of I/O options to deliver truly optimized systems for specific requirements. Users can optimize the storage and networking alternatives to accelerate performance, increase efficiency and find the perfect fit for their data center applications.
As Supermicro’s de facto standard multi-node architecture system for the most demanding data center and HPC applications, the flagship BigTwin reduces the complexity of compute clusters by doubling the density over the traditional single-node system. It supports four nodes with dual 3rd Gen AMD EPYC™ processors, up to 200G HDR networking through two PCI-E x16 and the flexible SIOM per node.
Building Block Solutions®
Supermicro provides a wide range of SKUs, and enables the company to build and deliver application-optimized solutions based upon customer requirements.
“Our building block architecture allows us to deliver a versatile portfolio of systems that maximize the benefits of the 3rd Gen EPYC Processors for specific workloads.” — Charles Liang, president and CEO.
The broadest portfolio of flexible and highly configurable systems deliver optimized performance, faster time to results, and better outcomes for every modern data center workload while providing cost-effective versatility with compelling TCO and TCE.