High Performance Computing
4 likes332 views
High performance computing (HPC) involves aggregating computing power for efficient parallel processing, essential for businesses needing to analyze large data volumes and solve complex problems. HPC is utilized across various industries such as aerospace, education, finance, and healthcare, offering benefits like reduced execution times, affordability via cloud solutions, and scalability. As a solution for computationally intensive challenges, HPC is crucial for enterprises, especially small and medium businesses, to compete effectively.
High Performance Computing
- 1. NOUSINFOSYSTEMS L E V E R A G I N G I N T E L L E C T HIGH PERFORMANCE COMPUTING (HPC): A BRIEF OVERVIEW Rajapandian | Senior Microsoft Consultant
- 2. 2High Performance Computing (HPC) ABSTRACT High Performance Computing refers to the activity of aggregating the computing power for parallel processing to run the advanced program efficiently. If a comput- er program takes a long time to complete or multiple instances of a computer program have to be run, then it’s the right time to choose HPC and utilize its power for reliable and quick results. BENEFITS OF HPC Most business entities need the ability to analyze the huge amount of data to solve their business challenges such as research, innovation of new products, increase revenue etc. HPC helps in the most complex calculations, simulation and modelling. HPC is affordable for businesses of all sizes, thanks to technical advancements and the cloud. HPC has grown tremendously in the last few years and is currently used in various sectors. OIL AND GAS H P C AEROSPACE & DEFENCE EDUCATION & RESEARCH EARTH SCIENCES FINANCIAL SERVICES LIFE SCIENCES MANUFACTURING
- 3. 3High Performance Computing (HPC) HPC OVERVIEW As we know, HPC uses parallel computing methodology with multiple computers to perform a common task. HPC is used to reduce the execution time, run multiple instances of a computer program and to meet larger memory needs and storage requirements of programs. In HPC, each computer/CPU is called as a node. When a CPU consists of multiple processors, each processor is referred to as a core. HPC cluster consists of a head node and compute nodes. Head node is responsible to allocate the task to compute nodes. Compute nodes are the set of loosely or tightly connected computers. Job is the set of tasks that the HPC has to execute. Each core will run a task from the job. User submits the job to head node through the user interface application. Multiple jobs can be submitted to the head node from different users at the same time. Depending on the availability of compute nodes, all the jobs can run at the same time or few jobs will wait in the queue for other jobs to complete. This queue based approach ensures that the hardware is utilized completely. Head node assigns the tasks to compute node from the job based on the priority of the jobs. User can monitor the status of the job using the tool in head node. Compute nodes process the data and write the result to the data server. The final result can be aggregated and viewed by the user interface application. U S E R I N T E R F A C E Job Submission HPC Cluster HPC Head Node HPC Compute Nodes Data Result
- 4. Head Node Computer Node Computer Node Computer NodeComputer Node Data Server Head Node Computer Node Computer Node Computer NodeComputer Node Data Server 4High Performance Computing (HPC) LOOSELY COUPLED SYSTEM The loosely coupled system can be chosen when the tasks in a job use different input parameters and the dependence between compute nodes is not needed. In loosely coupled system, each compute node doesn’t have the knowledge of the definitions of other compute nodes. TIGHTLY COUPLED SYSTEM Tightly coupled system is chosen when the job requires multiple compute nodes working together to get results. This needs the communication between the compute nodes and they depend on one another. Hence they are scalable to minimum extent only. HPC CLOUD The HPC requires multiple computers to run in parallel. Hence the investment on hardware will be high and it will increase as the business grows. There are chances that the HPC will not be used on day-to-day basis. The user may run the application using HPC once in a month or quarter. In this scenario, the investment done on the hardware is idle for a long time. HPC cloud comes on a pay-per-use basis, with an access to infinite resources. Cloud provides instant availability of systems with configuration requested by the user. The initial investment and maintenance cost in the cloud is less. Loosely coupled systems are a better fit for the cloud. Tightly coupled systems have some limitations on the cloud. Highly scalable applications make of use of the cloud to dynamically scale resources up and down as needed. A Private Cloud is the set of computers managed by a third party for the exclusive use of an enterprise. It is a highly controlled environment and configured behind the firewall. It focuses on compliance and security. A Public Cloud is the set of computers managed by a third party for use by other companies. A Hybrid cloud is the combination of a Private Cloud and a Public Cloud. In Hybrid Cloud, an enterprise can use private cloud for sensitive and critical workloads and public cloud for less sensitive information.
- 5. 5High Performance Computing (HPC) HPC IN AEROSPACE & DEFENCE HPC is used to design aircraft, missile and rocket using virtual simulation without spending millions of dollars in prototyping the product. HPC is also used for naviga- tion, surveillance and reconnaissance. It is used in computational structural mechan- ics, fluid dynamics and electrodynamics. Sometimes the data is gathered in real time and analyzed on the fly using HPC to provide the result to jet fighters immedi- ately. HPC IN EDUCATION AND RESEARCH Academic institutions educate the science and engineering students about computational science techniques to position them for careers in the cutting edge technology. Research and academic institutions use HPC to run calculations and simulations to address the grand challenges of computing such as develop- ing dynamic math models, design of control systems and simulation of electronic circuits or mechanical systems. HPC IN EARTH SCIENCES In earth sciences, models are developed to study the complex processes and interactions between the numerous components of the earth system. HPC uses ensemble modelling to simulate climate change for weather forecasting and predict seismic events etc., thus reducing the risk of environmental damage. It is used for short range weather forecasting, seasonal monsoon predic- tions and atmospheric chemistry model- ling.
- 6. 6High Performance Computing (HPC) HPC IN FINANCIAL SERVICES HPC has been used in the financial sector for many years. Banking and financial sectors use HPC to run thousands of calculations using stochastic modelling. This is to determine the risk at the enter- prise or customer level or actuarial analy- sis for insurance. HPC is also used in pricing and valuation of securities and derivatives. HPC helps financial institu- tions to respond quickly to the changing market environment and take necessary steps if risk arises. HPC IN MANUFACTURING HPC IN LIFE SCIENCES Life sciences is a highly R&D intensive sector. Computer simulation and model- ling is essential to study the complex biochemical systems. HPC is used by healthcare organizations and pharmaceu- tical companies to improve the effective- ness of clinical treatments. HPC is used in genomics workflows such as DNA and RNA sequence assembly and alignment. It is also used in proteomics. HPC is used in manufacturing sectors such as chemicals, food processing, glass, cement, paper, etc., to improve the efficiency of product development through Computer Aided Engineering and simulation. HPC uses predictive modelling in the manufacturing sector where it simu- lates and processes millions of data points to analyze material and design. In the automobile industry, a crash test is an expensive one. So manufacturers use computer modelling and simulation initial- ly to conduct these tests, before they do the actual crash test. HPC is used to improve safety features, reduce noise and vibration and the for simulations in wind tunnel experiments. HPC IN OIL & GAS The process of finding out the Oil or Gas drilling location requires significant computational power. Oil and gas organi- zations depend on HPC to process huge amounts of data provided by seismic acquisition systems and depth imaging software.
- 7. CONCLUSION HPC is essential to be a leader in most of the industries today. Many business challenges are highly computational-in- tensive. HPC is a helping hand to many sectors in overcoming these challenges efficiently. Large enterprises have their own in-house HPC cluster. HPC cloud attracts SMEs and SMBs to use the HPC without a huge investment and grow their business. ABOUT THE AUTHORS Rajapandian works as a Senior Microsoft Consultant with Nous Infosystems. He has an overall experience of 8 Years in Microsoft Technologies. His competency and areas of expertise include product and solution analysis and development. He has vast experience in Insurance and Music & Entertainment domains. He is a Computer Science B.Tech Graduate from National Institute of Technology, Tiruchirappalli. Rajapandian Senior Microsoft Consultant Nous Infosystems REFERENCES • http://insidehpc.com • http://blogs.msdn.com/b/amit_chatterjee/archive/2010/12/19/ parallelism-in-the-cloud-cluster-and-client.aspx • http://blogs.technet.com/b/windowshpc/archive/2008/04/14/ how-that-node-socket-core-thing-works.aspx • https://books.google.co.in/books?isbn=1439856427 • https://books.google.co.in/books?isbn=0387097651 ABOUT NOUS INFOSYSTEMS Nous Infosystems is a CMMi Level 5 SVC + SSD v1.3, ISO 9001:2008, and ISO/IEC 27001:2013 certified global Information Technology firm providing software solutions to the financial institutions across the globe. Major offerings include Digital Transformation, Application Development & Maintenance, Enterprise Application Integration, Product Engineer- ing, Business Intelligence, Independent Testing and Infrastructure Management Services.