HPC

What is HPC?

          

High Performance Computing (HPC) refers to the aggregation of computing power to provide greater computational performance than traditional computers or servers. High Performance Computing (HPC) or Supercomputing is just like everyday computing, the difference is only that it is more powerful. It helps people explore some of the world's biggest problems in science, engineering, and business by aggregating structures and using multiple computers and storage devices to process large amounts of data at very high speeds.

         

Currently, HPC is mostly used to solve performance-intensive, complex problems, and more and more organisations are moving their HPC workloads to the cloud. Cloud HPC will greatly optimise the economics of product research and development by reducing the need for prototypes, speeding up testing and reducing time to market.

            

How Does HPC Work?

         

In practice, there are workloads (e.g. DNA sequencing) that are too large for any single computer. HPC or supercomputing environments enable multiple nodes (computers) to work together in clusters (interconnected groups) to perform massive amounts of computation in a short period of time to meet the challenges of these large and extremely complex workloads. In addition, HPC reduces costs by supporting the automatic creation and deletion of clusters in the cloud.

         

HPC supports many types of workloads, two of the most common being parallel workloads and tightly coupled workloads.

          

Parallel Workloads

        

refers to computing problems that are subdivided into multiple small, simple, independent tasks that can run simultaneously, often with little to no communication with each other. For example, a business may submit 100 million credit card records to individual processor cores in a cluster of nodes. Processing a single credit card record is a small task, and when 100 million records are distributed across the cluster, 100 million small tasks can be executed simultaneously (in parallel) at an astonishing rate. Common usage scenarios for parallel workloads include risk simulation, molecular modelling, contextual search and logistics simulation.

             

Tightly Coupled Workloads

            

typically take up more shared resources and are broken down into smaller tasks that communicate with each other continuously. In other words, individual nodes in a cluster communicate with each other as they perform processing. Common usage scenarios for tightly coupled workloads include computational fluid dynamics, weather forecast modelling, materials simulation, automotive crash simulation, geospatial simulation and traffic management.

Why is HPC Important?

            

HPC has played an important role in academic research and industry innovation for decades. It has helped engineers, data scientists, designers, and other researchers solve many large and complex problems more economically and efficiently.

            

Key benefits of HPC include:

          

Reduced Physical Testing: HPC can be used to create simulations without the need for physical testing by the user. For example, when testing automotive accidents, HPC can generate crash simulations more economically and easily than actual crash tests.

          

Speed: Based on the latest CPUs, graphics processing units (GPUs) and low-latency network architectures such as RDMA (Remote Direct Memory Access), as well as all-flash local storage and block storage devices, HPC can perform massive computations in minutes that would have taken weeks in the past.

            

Cost: HPC can deliver answers faster and at a lower cost. And with cloud-based HPC, even small businesses and startups can easily run HPC workloads - paying only for the resources they actually use and scaling up and down as needed.

               

Innovation: HPC drives innovation in virtually every industry and is a key driver of breakthrough scientific discoveries that improve the quality of life for people around the world.

             

HPC Usage Scenarios 

        

HPC is becoming increasingly popular, with Fortune 1000 companies using HPC in almost every industry, and the global HPC market is expected to reach $44 billion by 2022, according to Hyperion Research.

        

Here are some of the industries where HPC is used and the corresponding HPC workload types:

             

Aerospace: Creating complex simulations, such as airflow over an aircraft wing.

              

Manufacturing: Enhancing the design, manufacture and testing of new products (e.g., autonomous driving) through simulation to produce safer cars and lighter parts, improve process efficiency and foster innovation.

                

Financial Technology (fintech): Performs complex risk analysis, high-frequency trading, financial modelling and fraud detection.

              

Genomics: Advancing genealogy research through DNA sequencing, drug-drug interaction analysis and protein analysis.

               

Healthcare: Developing drugs, developing vaccines, and researching innovative therapies for rare and common diseases.

                

Media and Entertainment: Creating animations, rendering film effects, transcoding large media files, and creating immersive entertainment experiences.

                       

Oil & Gas: Perform spatial analyses and test reservoir models to predict the location of oil and gas resources; simulate fluid flow and seismic data processing, etc.

Retail: analysing vast amounts of customer data to provide more targeted product recommendations and better service to customers.