News

Back

CSC status as an Intel® Parallel Computing Center renewed

To meet future scientific computing demands, computer systems in the next decade will have millions of processor cores.  To utilize these systems, today's codes need to be modernized. To support this process CSC, the Finnish center of expertise in ICT for research, education, culture, public administration and enterprises, has been granted a continuation of its status as an Intel® Parallel Computing Center (Intel® PCC).

The Intel PCC program provides funding to universities, institutions and research labs to modernize key community codes used across a wide range of disciplines to run on current state-of-the-art parallel architectures. The primary focus is to modernize applications to increase parallelism and scalability through optimizations that leverage cores, caches, threads, and vector capabilities of microprocessors and coprocessors.

"This grant enables us to significantly improve the support for modernization of scientific software and will directly benefit researchers that use our large-scale computing platforms," says Director Per Öster from CSC, the Principal Investigator in CSC's Intel® PCC project.

Successful cases in code optimization

CSC had a highly successful first term as an Intel Parallel Computing Center in 2014–2016, when several scientific software were modernized for platforms based on the Intel® Xeon® processor and Intel® Xeon Phi™  processor.

The close collaboration with Intel also resulted in CSC getting early access to a new generation Intel® Xeon Phi™ processors (formerly codenamed Knights Landing), which will be a key development platform for the next two years.

Sharing best practices

The upcoming Intel® PCC project will tune a larger number of codes for the latest architectures at three different levels of detail and intrusiveness:

  1. At the initial core level, best practices will be developed for compiling and running applications, and these recipes will be published as open source.
  2. Based on performance analysis a subset of applications will be further optimized by doing localized changes in the code, and the patches will be contributed to the original authors of the code.
  3. As a last step, a few applications will undergo an in depth optimization effort where the underlying data structures and algorithms will be rewritten.

"An important part of the project is to share knowledge. This work will be disseminated through courses, a dedicated web page and on GitHub," says Project Manager Sebastian von Alfthan.

Increased scientific productivity

The most obvious and visible outcomes of this work are a number of optimized scientific applications. 

"Serving as an Intel® PCC increases the knowledge in-house on how to extract maximum performance from modern Intel processors, and gives us insight into the performance of Intel Xeon processors and Xeon Phi processors, which aids in future updates to CSC's infrastructure," von Alfthan continues. 

According to the project manager, however, the most important goal is to support CSC's customers and increase their scientific productivity. Optimized applications and a higher level of computational expertise from the courses and white papers produced by the center may assist new scientific breakthroughs.

Additional information

github.com/cschpc

Per Öster
Director, Research Infrastructures
Principal Investigator of Intel® PCC at CSC
per.oster@csc.fi
+358503819030

Sebastian von Alfthan
Senior Applications Specialist
Project Manager of Intel® PCC at CSC
sebastian.von.alfthan@csc.fi
+358405888688

 

Read more about the results of CSC's Intel® PCC project 2014-2016 on CSC Blog:
Code modernization: 4 successful cases and lessons learned
https://www.csc.fi/web/blog/post/-/blogs/code-modernization-4-successful-cases-and-lessons-learned

 

Intel, Xeon, and Xeon Phi are trademarks or registered trademarks of Intel Corporation in the United States and other countries.