Based on Elastic Bare Metal (EBM) instances and high-speed interconnectivity of RDMA (Remote Direct Memory Access) technology, Super Computing Cluster (SCC) servers improve network performance and increase the acceleration ratio of large-scale clusters. SCC servers boast all the advantages of EBM instances and feature high-quality network performance with high bandwidth and low latency.
SCC servers cater to high-performance computing for artificial intelligence, science/engineering computing, and audio/video processing.
- High-speed RDMA Connection
Supports high-speed RDMA network connection and GPU-direct to secure high bandwidth and low latency.
- Ultimate Computing Performance
Deploy independent CPU and GPU Elastic Bare Mental servers with ultra-high computing capabilities in each single node.
Elastic Bare Mental servers capable of deploying in minutes and quick cluster resizing as needed.
Highly secure isolation of physical servers to meet high availability and data security requirements.
SCC offers up to 96 CPU vCores (Skylake architecture), up to 512GB of memory, and up to 8 GPU cards (Pascal or Volta architecture) within a single instance to meet the most demanding of data processing power requirements for large-scale applications
Establishes the latest in advanced network architecture through RoCE v2, enabling RDMA interconnection, significantly improved cluster acceleration, and the ability to break through network bottlenecks
Create a supercomputing cluster with one click and no queueing. You can also automatically increase or decrease cluster nodes without the limit of a physical machine and enjoy flexible Pay-As-You-Go billing.
High Speedup Computing Cluster
Boost an interconnection through the latest in network architecture
Deploy powerful instances built on advanced hardware such as the Intel Skylake CPU, RoCE v2 RDMA network and Nvidia P100/V100 GPU. The superior acceleration achieved by high-speed networking maximizes the performance of multiple instance clusters.
High speed data communication
RDMA network enables high-speed communication of large-scale data such as GPUDirect, which significantly enhances multi-node cluster acceleration.
Superior computing performance
Meet advanced computing power needs for big data + AI applications by combining the power of CPU and GPU in a single architecture.
Interconnectivity of Alibaba Cloud services
SCC seamlessly integrates with other Alibaba Cloud products and services to provide various solutions and serve different requirements.