In this video from SC18 in Dallas, Prasad Rampalli describes how Accenture is taking HPC to the Enterprise.
The Accenture Engineering Compute (AEC) digital platform is a first mover in the industry with a highly differentiated AIML/analytics based Job-resource prediction and orchestration engine to ensure the most efficient placement of HPC workloads on-prem or in the cloud. The solution delivers dramatic gains in capital equipment utilization (~2x) , job throughput (~30%) and business agility with automated cloud bursting for HPC/Grid computing environments. It also includes fine-grain demand-capacity planning and an optimized data center rack & storage design as part of the integrated software offering. The platform is fully extensible to varied grid scheduling environments and typically demonstrates results within 180 days from the on-set of a POC. Dense compute clusters ranging from 10,000- 60,000 nodes can potentially realize ~ $20m – $100+m in capx savings over 3 years. This solution is being successfully deployed across one of the largest and most complex global grid environments for a leading semiconductor company with others in value assessment/POC discussions.”
