Micro Focus is now part of OpenText. Learn more >

You are here

You are here

Why big data makes modernizing code mandatory

public://pictures/Suresh Aswani.png
Suresh Aswani Senior Manager, Solutions Marketing, Hewlett Packard Enterprise
 

We live in an era of big data, and with big data comes big ideas—big ideas that help you succeed, that make people's lives better, that literally change the world.

But your valuable raw data doesn't transform into useful information on its own. It needs powerful and precise machines to handle it effectively. It needs modern processors with multiple cores. It needs multilevel parallelism. What you—and your data—need is high-performance computing (HPC) capabilities.

Most of all big data means modernizing your code. This is an urgent matter for data-oriented organizations. Make 2016 the year you tackle it.

Why you should modernize your code

HPC is all about doing really complex things really well and really fast. Any data-oriented organization trying to stay competitive needs infrastructure that can process large amounts of information extremely quickly.

Modern processors have multiple cores, which allow multiple threads of code to run in parallel. When modernized code is run on HPC machines, complementary sets of instructions are being executed simultaneously on different cores. This strategy greatly increases the power and speed of computing, opening up a range of opportunities for new and enhanced applications. Multicore CPUs, with their attendant multithreading, enable better performance and better throughput on nearly any application.

And yet, despite the possibility of greatly improved performance, many organizations still run single-thread program codes, which don't take full advantage of today's powerful computing platforms. They either haven't modernized their infrastructure at all or they're running single-thread code on multicore computers.

Running outdated codes on powerful modern processors is like propelling a motorboat with a paddle: You'll get where you're going but slowly and with a lot of unnecessary effort—while being rocked by the wake that your competitors' speedboats leave behind. All the while, you could be using serious horsepower and precision steering to get you to your destination. All you need is modernized code to fire things up.

Big ideas are executed by modernized code

HPC is transforming all kinds of industries, with code modernization helping to dramatically alter their landscapes. Consider high-frequency financial trading, where every millisecond matters. Or think about health and life sciences, where genome- and DNA-related computer applications may save lives. Or oil and gas, where exploration is difficult and costly. The list goes on.

The complex work of sequencing genomes used to take days, but it can now be done much faster using modern processors running multithreaded code. Seismic processing for oil and gas exploration, which typically requires heavily graphical applications, can now be done much more easily with the powerful capabilities of HPC and modernized code.

Another great example of how HPC transforms industries is the smart energy meter data analytics system. Energy use varies from hour to hour, falling and peaking during certain times of day and night. Smart meters record data on these usage patterns, and an analytics system uses parallelism and HPC software to assess the data in real time and then forecast consumption across geographic areas. The analysis allows for more effective and sustainable distribution of energy among the cities on the grid.

What kind of new application or process could your industry benefit from if your computing infrastructure could handle the task? The possibilities are endless. You just need the right tools.

How you can modernize your code

Any data-oriented organization will see sooner or later that code modernization is inevitable. The only question is, how do you get started?

A number of great tools are available. HPE is working jointly with Intel to utilize tools like the Intel Parallel Studio XE, which helps developers write multithreaded code faster. High-performance compilers and multicode debuggers, among other useful tools, are becoming increasingly available, but many developers may not know how to use these tools.

That's where we come in. HPE and Intel jointly run HPC Centers of Excellence (CoE) in Houston and Grenoble, France, where you may qualify to take advantage of the infrastructure, tools, and expert advice needed to modernize your code and join the world of HPC. We have test servers with multicore processors, a library of tools like those discussed above, and a roster of people to help you figure out your next steps and execute your code modernization plan.

To work with us at an HPC CoE, first get in touch with your HPE contact person, who will connect you to a CoE. Most of the time it’s best to work on-site at the CoE. Once you're there, a project manager will work with you to assess your code modernization needs, decide which experts will help you most, and lay out a plan for the work.

The biggest challenge to modernizing your code

Leveraging the performance improvements enabled by HPC will allow you to find new ways to confront the real-life, complex problems that confront your industry. Developers know this, but administrators sometimes need to be convinced.

Some organizational leaders need to be inspired to embrace both the new machines and the new code that fit those servers best. Others are already investing in multicore processors, but need to be enlightened on how modernized code is also necessary to make that investment worthwhile.

Unfortunately, convincing anyone of the wonders of a new technology is difficult if you don't have the tools to demonstrate its benefits. Even when organizational leadership understands that these modern processors can help their applications run faster, they want to see the potential performance improvement clearly spelled out. They need to know that the return on investment will be high enough to justify the investment of money and time.

With HPE, you'll be able to pretest your current code and applications on the most cutting-edge machines, then modernize and test portions of your code to compare performance against the unaltered code. Documenting the difference in performance will help you justify to higher-ups that the company needs this kind of updated infrastructure—both multicore processors and a full modernization of your code base.

The results of preliminary testing on these tools will show clearly how modernization can help your company compete in a world of rapid technological change. Big data + modern infrastructure = big ideas. Get out there and succeed. Change the world. This year, choose to put away the paddle and turn on your engine. 

Keep learning

Read more articles about: Enterprise ITData Centers