Graphcore systems excel at both training and inference. The highly parallel computational resources together with graph software tools and libraries, allows researchers to explore machine intelligence across a much broader front than current solutions. Graphcore technology allows recent successes in deep learning to be effectively and efficiently applied towards algorithms developed for general artificial intelligence, NLP, financial risk analysis and more.
Graphcloud enables customers to use IPU-POD16 and IPU-POD64 configurations as a service for a unique solution for massive, disaggregated scale-out enabling high-performance machine intelligence compute to supercomputing scale. The IPU-POD64 builds upon the innovative IPU-M2000 and offers seamless scale-out to work as one integral whole or as independent subdivided partitions to handle multiple workloads and different users.
The IPU-POD64 utilized in Graphcloud consists of 16 IPU-M2000s. IPU-PODs communicate with near-zero latency using Graphcore's unique IPU-Fabric™ interconnect architecture. IPU-Fabric has been specifically designed to eliminate communication bottlenecks and allow thousands of IPUs to operate on machine intelligence workloads as a single, high-performance and ultra-fast cohesive unit.
With Poplar, managing IPUs at scale is as simple as programming a single device, allowing you to focus on the data and the results. Graphcore's state of the art compiler simplifies IPU programming by handling all the scheduling and work partitioning of large models, including memory control, while the Graph Engine builds the runtime to execute your workload efficiently in Graphcloud. Graphcore has made it possible to dynamically share your AI compute, with Graphcore’s Virtual IPU software. You can have tens, hundreds, even thousands of IPUs working together on model training.
Graphcore's Poplar software supports standard ML frameworks including TensorFlow, PyTorch, ONNX and PaddlePaddle as well as industry standard converged infrastructure management tools so it’s easy to deploy, including Open BMC, Redfish, Docker containers, and orchestration with Slurm and Kubernetes.
Cloud security means protecting a complex web of workloads consisting of hardware, applications, and data spread across edge, core, remote workforce, and cloud deployments. Cirrascale Cloud Services’ data centers and infrastructure are architected to protect your information, identities, applications, and devices.
Accessing Graphcloud is secure and easy. Cirrascale uses industry-standard link encryption and authentication methods such as IPsec and SSH to ensure the security of data during transmission. A multitiered set of defensive services are also used to strictly control network access. Your data remains safe throughout your time using our service. Once you opt to exit our service, our engineering operations team personally verifies that your data has been removed.
Customers can utilize Megaport connectivity through Cirrascale and an AWS DirectConnect, Azure ExpressRoute, or Google Cloud Partner Interconnect to enable a seamless connection between a customer’s VPC at these hyperscalers and their Graphcloud instance.
This connection lets customers connect to any one of, or to a multiple of, the world’s leading hyperscalers and access best-of-breed services. Additionally, it enables customers to connect to multiple cloud regions from a single interconnection point with greater security and privacy than traditional networking solutions.
Using a Graphcore IPU cloud instance with Cirrascale ensures no hidden fees with our flat-rate billing model. You pay one price without the worry of fluctuating bills like those at other providers. Pricing shown below is per instance for the total time stated.
|IPU Instance||IPU Specs||Weekly Rate||Monthly Rate||1-Year Rate||3-Year Rate|
|IPU-POD 16||4x IPU-M2000||$5,000||$16,000
|IPU-POD 64||16x IPU-M2000||$20,000||$64,000
Cirrascale has partnered with the industry’s top storage vendors to supply our customers with the absolute fastest storage options available. Connected with up to 100Gb Ethernet, our specialized NVMe hot-tier storage offerings deliver the performance needed to eliminate workflow bottlenecks.
|NVMe Hot-Tier Storage||<50TB||$0.40|
|NVMe Hot-Tier Storage||>50TB||$0.20|