Capable of tackling the world's biggest AI challenges.
Faster I/O - up to 5.6x more I/O bandwidth than x86 servers
The Power AC922 includes a variety of next-generation I/O architectures, including PCIe Gen4, CAPI 2.0, OpenCAPI and NVIDIA(R) NVLINK(TM), to provide up to 5.6 times as much bandwidth for data-intensive workloads versus compared x86 servers.
PCIe Gen4 – 2x the bandwidth of PCIe Gen3
Gen3The Power AC922 is the industry's first server to feature the next generation of the industry standard PCIe interconnect. PCIe generation 4 delivers approximately 2x the data bandwidth of the PCIe generation 3 interconnect found in x86 servers.
POWER9 processor—the latest POWER processor, designed for AI
Built for the AI era, POWER9 supports up to 5.6x more I/O and 2x more threads than its x86 contemporaries. POWER9 is available on configurations with anywhere between 16 and up to 44 cores in the Power AC922 server.
Advanced GPUs - up to 6 NVIDIA® Tesla® V100 GPUs with NVLink
The Power AC922 pairs IBM POWER9 CPUs and NVIDIA Tesla V100 with NVLink GPUs, creating a server capable of delivering up to 5.6x times the performance. This provides massive throughput capability for HPC, deep learning and AI workloads.
Coherence - share RAM across CPUs & GPUs
CPU to GPU coherence in the Power AC922 addresses growth concerns by allowing accelerated applications to leverage system memory as GPU memory. This coherence helps simplify programming by reducing data movement and locality requirements.
Built scalable from one server to supercomputer
The Power AC922 is the backbone of the Department of Energy's Summit and Sierra supercomputers. With its scalability, efficiency and ease of AI deployment, it's also ideally suited to address your organization's AI aspirations.