Seattle, WA, USA and Barcelona, Spain - December 7, 2016 - At the 2016 Neural Information Processing Systems (NIPS) Conference in Barcelona, Spain, global supercomputer leader Cray Inc. (Nasdaq: CRAY) today announced the results of a deep learning collaboration between Cray, Microsoft, and the Swiss National Supercomputing Centre (CSCS) that expands the horizons of running deep learning algorithms at scale using the power of Cray supercomputers.
|
Swiss National Supercomputing Centre (CSCS) in Lugano, Switzerland.
Main entrance in the office building.
Photo courtesy of CSCS |
|
Seattle, WA, USA and Barcelona, Spain – December 7, 2016
Running larger deep learning models is a path to new scientific possibilities, but conventional systems and architectures limit the problems that can be addressed, as models take too long to train.
Cray worked with Microsoft and CSCS, a world-class scientific computing center, to leverage their decades of high performance computing expertise to profoundly scale the Microsoft Cognitive Toolkit (formerly CNTK) on a Cray® XC50™ supercomputer at CSCS nicknamed “Piz Daint”.
|
Swiss National Supercomputing Centre (CSCS) in Lugano, Switzerland.
The office building (right) is connected on the Northern side to the computer building (left) by a bridge and an underground tunnel.
Photo courtesy of CSCS |
|
By accelerating the training process, instead of waiting weeks or months for results, data scientists can obtain results within hours or even minutes.
With the introduction of supercomputing architectures and technologies to deep learning frameworks, customers now have the ability to solve a whole new class of problems, such as moving from image recognition to video recognition, and from simple speech recognition to natural language processing with context.
Deep learning problems share algorithmic similarities with applications traditionally run on a massively parallel supercomputer.
|
The Cray XC30 supercomputer “Piz Daint” at CSCS in Lugano, Switzerland.
Photo courtesy of CSCS |
|
By optimizing inter-node communication using the Cray® XC™ Aries network and a high performance MPI library, each training job can leverage significantly more compute resources – reducing the time required to train an individual model.
|
Prof. Dr. Thomas C. Shulthess, director of the Swiss National Supercomputing Centre (CSCS).
Prof. Thomas Schulthess showing a computing blade of the Cray XC30 supercomputer „Piz Daint” - September 12, 2013.
Photo courtesy of CSCS |
|
“Cray’s proficiency in performance analysis and profiling, combined with the unique architecture of the XC systems, allowed us to bring deep learning problems to our Piz Daint system and scale them in a way that nobody else has,” said Prof. Dr. Thomas C. Schulthess, director of the Swiss National Supercomputing Centre (CSCS).
“What is most exciting is that our researchers and scientists will now be able to use our existing Cray XC supercomputer to take on a new class of deep learning problems that were previously infeasible.”
|
Dr. Xuedong Huang, distinguished engineer, Microsoft AI and Research.
Photo courtesy of LinkedIn |
|
“Applying a supercomputing approach to optimize deep learning workloads represents a powerful breakthrough for training and evaluating deep learning algorithms at scale,” said Dr. Xuedong Huang, distinguished engineer, Microsoft AI and Research.
“Our collaboration with Cray and CSCS has demonstrated how the Microsoft Cognitive Toolkit can be used to push the boundaries of deep learning.”
|
Back view of the Cray XC30 supercomputer „Piz Daint” at CSCS in Lugano, Switzerland.
Photo courtesy of CSCS |
|
A team of experts from Cray, Microsoft, and CSCS have scaled the Microsoft Cognitive Toolkit to more than 1,000 NVIDIA® Tesla® P100 GPU accelerators on the Cray XC50 supercomputer at CSCS.
The result of this deep learning collaboration opens the door for researchers to run larger, more complex, and multi-layered deep learning workloads at scale, harnessing the performance of a Cray supercomputer.
To simplify the building and deploying of deep learning environments in supercomputing, Cray is supporting its Cray XC customers with deep learning toolkits, such as the Microsoft Cognitive Toolkit, that allow customers to run deep learning applications at their fullest potential – at scale on a Cray supercomputer.
Fusing high performance computing capability with deep learning is another step forward in Cray’s vision of the convergence of supercomputing and big data.
|
The Cray XC40 system offers the combined advantages of the Aries interconnect and Dragonfly network topology, multi-core and many-core processors, integrated I/O acceleration options and Cray OS and programming environment, delivering up to 100 PF sustained system performance.
It’s designed for production supercomputing and user productivity.
Photo courtesy of Cray |
|
“Only Cray can bring the combination of supercomputing technologies, supercomputing best practices, and expertise in performance optimization to scale deep learning problems,” said Dr. Mark S. Staveley, Cray’s director of deep learning and machine learning.
“We are working to unlock possibilities around new approaches and model sizes, turning the dreams and theories of scientists into something real that they can explore. Our collaboration with Microsoft and CSCS is a game changer for what can be accomplished using deep learning.”
About Cray Inc.
Global supercomputing leader Cray Inc. (Nasdaq:CRAY) provides innovative systems and solutions enabling scientists and engineers in industry, academia and government to meet existing and future simulation and analytics challenges.
Leveraging more than 40 years of experience in developing and servicing the world’s most advanced supercomputers, Cray offers a comprehensive portfolio of supercomputers and big data storage and analytics solutions delivering unrivaled performance, efficiency and scalability.
Cray’s Adaptive Supercomputing vision is focused on delivering innovative next-generation products that integrate diverse processing technologies into a unified architecture, allowing customers to meet the market’s continued demand for realized performance.
Go to www.cray.com for more information.
Find here more information on Cray’s machine learning and deep learning solutions
http://www.cray.com/solutions/machine-learning-deep-learning
and the Cray XC series of supercomputers.
http://www.cray.com/products/computing/xc-series/
Cray Media:
Nick Davis
206/701-2123
pr@cray.com
Cray Investors:
Paul Hiemstra
206/701-2044
ir@cray.com
Sources:
Swiss National Supercomputing Centre (CSCS)
http://www.cscs.ch/index.php
CRAY
http://www.cray.com
ASTROMAN Magazine - 2016.11.09
Intel's Vision: Smart and Connected to the Cloud
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2154
ASTROMAN Magazine - 2016.10.16
Tech Leaders Unite to Enable New Cloud Datacenter Server Designs for Big Data, Machine Learning, Analytics, and other Emerging Workloads, called OpenCAPI
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2138
ASTROMAN Magazine – 2016.10.01
Amazon, DeepMind/Google, Facebook, IBM and Microsoft Establish Partnership on Artificial Intelligence Best Practices
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2131
ASTROMAN Magazine – 2016.10.01
AUDI AG, BMW Group, Daimler AG, Ericsson, Huawei, Intel, Nokia and Qualcomm form global cross-industry 5G Automotive Association
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2132
ASTROMAN Magazine - 2016.10.01
IBM Unveils Industry’s First Platform to Integrate All Data Types for AI-Powered Decision-Making
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2130
ASTROMAN Magazine - 2016.09.03
IFA 2016: IBM Watson Powers Wave of Innovation in Consumer Electronics
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2115
ASTROMAN Magazine - 2016.08.27
Piton: A 25-core New Academic Manycore Research Processor, Demonstrates Efficiency and Scalable Design at Hot Chips by Princeton University Researchers
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2112
ASTROMAN Magazine - 2016.08.12
Intel: The Foundation of Artificial Intelligence
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2102
ASTROMAN Magazine - 2016.07.30
NVIDIA: What's the Difference Between Artificial Intelligence, Machine Learning, and Deep Learning?
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2097
ASTROMAN Magazine - 2016.07.15
Microsoft Worldwide Partner Conference 2016
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2088
ASTROMAN Magazine – 2015.12.12
Intel, Vodafone, Orange, Deutsche Telekom, Ericsson, Nokia and more speaking at The Internet of Things Tech Expo Europe, 10-11th February 2016
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=2002
ASTROMAN Magazine - 2015.08.29
IBM and GENCI Team to Drive Supercomputing Closer to Exascale
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=1952
ASTROMAN Magazine - 2015.08.09
IBM: Watson to Gain Ability to "See" with Planned USD1 Billion Acquisition of Merge Healthcare
http://www.astroman.com.pl/index.php?mod=magazine&a=read&id=1946
Editor-in-Chief of ASTROMAN magazine: Roman Wojtala, Ph.D.
|