The Trusted Leader in High Performance Computing


Supercomputing Accelerates Path to Cancer Predictions and Treatment

January 10, 2017

Quantitative bioinformatics studies utilizing genomics, proteomics and imaging experiments require the integration of vast amounts of data in order to gain new insights from large population data sets. In an effort to solve some of life science's most complex questions, The Linding Laboratory in Denmark partnered with HPE/SGI to further its research at the Biotech Research and Innovation Centre (BRIC), University of Copenhagen and extract deeper insights into areas such as cell and cancer biology using high performance computing (HPC).

Understanding the Disease

Much of the study focuses on disease research, including understanding how and why disease occurs and discovering cancer-prone genes to identify new diagnostic markers. These markers can then be used to predict how cells will respond to cancer treatment and determine the most effective treatment for each individual patient.

High Performance In-Memory Computing and the Road to Better Treatments

BRIC researchers turned to HPE/SGI to develop a supercomputing solution to manage the massive volumes of data associated with this type of research. The research organization highlights the importance of big data in cancer biology and underpins the necessity for large dynamic-range computing platforms to conduct its work.

With the HPE/SGI® UV™ 300, the University of Copenhagen Is Enabled to Develop Better Treatments for Cancer Patients and Improve Cancer Research

The university deployed an HPE/SGI® UV™ 300 machine for its unique capabilities, including large shared memory, unmatched data performance and a single OS. With 7 Terabytes of in-memory computing power, the powerful HPE/SGI machine quickly provides researchers with new insights from its data. The new system's algorithms are aimed to quickly and accurately forecast cell behavior, making it possible for researchers to identify a speedier path towards research breakthroughs.

"The identification of distinct changes within our tissues that help predict cancer progression is a major step forward, and we are confident it can aid in the development of novel therapies and screening techniques," says Dr. Rune Linding, professor of cellular signaling and research group leader for the Linding Laboratory at the University of Copenhagen. "In these studies we have generated over a trillion data points that our computational algorithms then use to forecast the behavior of perturbed cancer cells, in a manner similar to aircraft or weather models. This is a vast computational and big data challenge that requires an extreme degree of computational flexibility. We have been surprised at the exceptional speed of the HPE/SGI UV 300. The system is really an order of magnitude faster. What is advantageous is that it combines the Intel® Xeon® Core Processor with Flash and Intel® Xeon Phi™ Processors. It is a very flexible system that allows us to do many things. The UV shared memory system made it possible to directly explore new creative ideas and swiftly follow a path towards a major breakthrough study. The ease of programming in a shared memory environment allows program threads to access the full datasets and lets the scientist focus on the problem at hand and not invest time in the technicalities of programming a distributed memory system."

Furthering Discovery with Compute Availability

The HPE/SGI UV 300 made it easier to program big memory applications since this type of research requires very large memory footprints. Previously, these researchers needed to consider HPC resources before they could conduct new research, which created a barrier to creativity and gleaning new research insights. The HPE/SGI system unleashed new creative energy at the institution as scientists can now explore new research angles without worrying about computing availability.

To manage the large amounts of data, the new UV system is equipped with HPE/SGI® DMF™ storage software, HPE/SGI's proven management of tiered data, which makes access to important data very fast while managing the total cost of data storage across multiple tiers. HPE/SGI's DMF allows researchers to manage data from a primary tier storage system that is backed up to a secondary tier using automated policies if files are not used for a certain amount of time and thus saves dollars for storage costs. This enables researchers to benefit from ultra-high bandwidth and IOPS (input/output operations per second), while holding all the data in fast, accessible storage to explore the many-many relationship.

"Cancer research requires continuous processing of massive data workloads," said Gabriel Broner, vice president and general manager of HPC, HPE/SGI. "It is essential to keep up with new behaviors of cancer cells as they become resistant to medicine. By using our HPE/SGI UV 300 supercomputer, BRIC is able to advance research results in multiple experiments. We are dedicated to advancing genomics and imaging technology to fill the gap between genetic diseases and their cures."

HPE/SGI sees tremendous growth opportunities in life sciences as medical research continues to advance and require powerful analysis. The company is looking forward to delivering HPC machines to these customers to enable breakthroughs in treatments that will ultimately create a world with less incurable diseases.

Announcement Highlights

  • The studies at the University of Copenhagen and Biotech Research & Innovation Centre (BRIC) highlight the importance of big data in cancer biology and underpin the importance of large dynamic-range computing platforms such as the HPE/SGI UV.
  • The HPE/SGI UV server platform offers unique capabilities for research computing, combining industry-leading shared-memory designs with unmatched data performance capabilities, making it an ideal choice for big data research workflows.
  • BRIC focuses on disease research for cancer, including understanding how and why disease occurs and discovering disease-related genes to identify new diagnostic markers and predict how cells will respond to cancer treatment. These markers can then be used to determine the most effective treatment for each individual patient.

Technical Information

  • HPE/SGI UV® 300™ system which includes:
  • Three HPE/SGI® UV™ 300 supercomputers with Numalink®7 interconnect that are 12 sockets each with 3GHz 12 core E7-8857 v2 7 TB in-memory compute power
  • Four Intel® Xeon® Phi™ processors and twelve Intel® Xeon® E7-8857 processors
  • Two Intel® Flash PCIe* NVMe P3700 solid-state drive series for a total of 1.6TB (Intel® SSD Data Center P3700 Series 1.6TB)
  • HPE/SGI® DMF™ storage software
  • HPE/SGI InfiniteStorage™ 5100 storage
  • The system also uses Rack Management Controller, SUSE Linux Enterprise Server and HPE/SGI® Performance Suite

© 2017 Silicon Graphics International Corp. All rights reserved. SGI and the SGI logo are trademarks or registered trademarks of Silicon Graphics International Corp. or its subsidiaries in the United States and/or other countries. Intel, Xeon, Xeon Core and Xeon Phi are trademarks or registered trademarks of Intel Corporation. All other product and service names mentioned are the trademarks of their respective companies.