1.what is HPC?
High performance computing (HPC) refers to computing systems and environments that usually use many processors (as part of a single machine) or several computers organized in a cluster (operating as a single computing resource). High Performance Computing (High Performance Computing) is a branch of computer science, which mainly refers to the research and development of high performance computers from the aspects of architecture, parallel algorithms, and software development. There are many types of HPC systems, ranging from large clusters of standard computers to highly specialized hardware. Most cluster-based HPC systems use high-performance network interconnections. HPC systems use specialized operating systems that are designed to look like a single computing resource. HPC solutions are also dedicated units that are specifically designed and deployed to act as (and only serve as) large computing resources. HPC solutions are designed to provide specific resource solutions, such as powerful computing power and storing large amounts of data in memory for The ability to handle them, the dedicated features of HPC solutions provide some benefits when developing applications to use this ability.
2.FP5295G2 came into being
With the rapid development of computer technology, the calculation speed of high-performance computers continues to increase, and its standards are also constantly changing. Most HPC systems present themselves as a single computing resource, so it becomes a programming responsibility, requiring a dedicated library to build an application that can be distributed to the entire resource. Application development in the HPC environment is usually handled through a dedicated library, which greatly simplifies the process of creating an application and allocating the tasks of the application to the entire HPC system. So Inspur created a product system designed for deep learning (DL) and artificial intelligence (AI), high-performance data analysis (HPDA) and high-performance computing (HPC). FP5295G2 is a next-generation product based on the POWER9 processor System, enterprises can confidently deploy data-intensive workloads such as deep learning frameworks and accelerated databases.
3.The functional characteristics of FP5295G2
1) Brand new AI infrastructure: CPU-GPU NVLink interconnection, shared memory The new AI infrastructure provides efficient support for enterprise data-intensive workloads. POWER9 CPU and NVIDIA Tesla V100 GPU are interconnected through NVLink, the only one that can provide up to 5.6 times the performance acceleration between CPU and GPU. NVLink acceleration supports CPU-GPU shared memory, simplifies the programming process, and enables customers to enjoy the high-performance experience of GPU. Provides richer and more efficient IO support, PCIe Gen4, CAPI 2.0, OpenCAPI and many other features. The total bandwidth that can be provided is 2-5.6 times that of PCIe Gen3 in x86 servers, with lower latency and faster response.
2) Joint practice of the best CPU and best GPU designed for the AI era
The POWER9 CPU specifically designed for AI, twice the number of x86 threads, larger L3 cache, NVLink native direct connection, and ultra-fast running speed, fully mobilize all the surrounding performance. It is equipped with 4 NVIDIA® Tesla® V100 GPUs that support NVLink technology, supports 32G of video memory, and a single GPU provides double-precision 15.7 TFLOPS ultra-high computing power. FP5295G2 single node provides Tensor performance of more than 500 trillion times per second, and you can even form them into a huge cluster of 3 trillion trillion AI operations per second as a service.
3) Enterprise-level AI platform is ready
The enterprise-level Power AI DL framework simplifies the deployment of deep learning in performance, and provides AI users with a more powerful and simple end-to-end tool chain. You can start with one node and efficiently scale to stacks or thousands of nodes, and their performance grows almost linearly.
4. Technical specifications of FP5295G2
FP5295G2 is divided into a standard version and an enhanced version. The difference between them is that the processor and GPU of the enhanced version increase performance on the basis of the standard version. For example, the standard version of the processor uses 2 POWER9 Monza processors with 16 cores with NVLink, while the enhanced version uses 20 cores. The standard version of the GPU is 4 x 16GB NVIDIA Tesla V100 with NVLink, and the enhanced version of the GPU is 4 x 16GB. The processor to memory bandwidth is 170GB/s per socket and 340GB/s per system. The PCIe expansion slot adopts 4 PCIe Gen4 slots. The I/O interface is 2 x USB 3.0, 2 x 1 GB network port, VGA interface. FP5295G2 carries two 2.5-inch SATA HDD/SSD hard disks, and PCIe NVMe SSD is optional. Optional support RAID 0/1/5/6/10/50/60, support Cache super capacitor protection, and support external USB optical drive. The operating system is Red Hat Enterprise Linux, Ubuntu Server, won the bid of Kylin, and Red Flag. Equipped with 1+1 redundant platinum power supply, the weight is about 30 kg, and the working environment temperature is 5℃-35℃.
FP5295G2 can help you deploy AI easily and efficiently, and is your ideal choice to realize your AI vision.