The Galactos Project: Using HPC To Run One of Cosmology's Hardest Challenges
Debbie Bard from NERSC gave this talk at the HPC User Forum. "We present Galactos, a high performance implementation of a novel, O(N^2 ) algorithm that uses a load-balanced k-d tree and spherical harmonic expansions to compute the anisotropic 3PCF. Our implementation is optimized for the Intel Xeon Phi architecture, exploiting SIMD parallelism, instruction and thread concurrency, and signicant L1 and L2 cache reuse, reaching 39% of peak performance on a single node. Galactos scales to the full Cori system, achieving 9.8 PF (peak) and 5.06 PF (sustained) across 9636 nodes, making the 3PCF easily computable for all galaxies in the observable universe." [READ MORE...]
Featured Stories
Podcast: Supercomputing the Emergence of Material Behavior
In this TACC Podcast, Chemists at the University of California, San Diego describe how they used supercomputing to design a sheet of proteins that toggle between different states of porosity and density. This is a first in biomolecular design that combined experimental studies with computation done on supercomputers. "To meet these and other computational challenges, Paesani has been awarded supercomputer allocations through XSEDE, the Extreme Science and Engineering Discovery Environment, funded by the National Science Foundation." [READ MORE...]
New Tachyum Prodigy Chip has "More than 10x the Performance of Conventional Processors"
Today Silicon Valley startup Tachyum Inc. unveiled its new processor family – codenamed “Prodigy” – that combines the advantages of CPUs with GP-GPUs, and specialized AI chips in a single universal processor platform. According to Tachyum, the new chip has "ten times the processing power per watt" and is capable of running the world’s most complex compute tasks. "With its disruptive architecture, Prodigy will enable a super-computational system for real-time full capacity human brain neural network simulation by 2020." [READ MORE...]
Satoshi Matsuoka to receive High Performance Parallel Distributed Computation Achievement Award
Today ACM announced that that Dr. Satoshi Matsuoka will receive the annual HPDC Achievement Award for his pioneering research in the design, implementation, and application of high performance systems and software tools for parallel and distributed systems. "ACM HPDC is one of the top international conferences in the field of Computer Science / High Performance Calculation, and among them, I am delighted to have won the Society Career Award for the first time as a Japanese." [READ MORE...]
Featured Resource
DNN Implementation, Optimization, and Challenges
This is the third in a five-part series that explores the potential of unified deep learning with CPU, GPU and FGPA technologies. This post explores DNN implementation, optimization and challenges. [READ MORE...]
Industry Perspectives
Video: Doug Kothe Looks Ahead at The Exascale Computing Project
In this video, Doug Kothe from ORNl provides an update on the Exascale Computing Project. "With respect to progress, marrying high-risk exploratory and high-return R&D with formal project management is a formidable challenge. In January, through what is called DOE’s Independent Project Review, or IPR, process, we learned that we can indeed meet that challenge in a way that allows us to drive hard with a sense of urgency and still deliver on the essential products and solutions. In short, we passed the review with flying colors—and what’s especially encouraging is that the feedback we received tells us what we can do to improve." [Read More...]
BioScience gets a Boost Through Order-of-Magnitude Computing Gains
Intel is working with leaders in the field to eliminate today’s data processing bottlenecks. In this guest post from Intel, the company explores how BioScience is getting a leg up from order-of-magnitude computing progress. "Intel’s framework is designed to make HPC simpler, more affordable, and more powerful." [Read More...]
Featured from insideBIGDATA
- Parallel Storage Fuels Groundbreaking Neuroscience and Behavioral Research at HarvardTo alleviate bottlenecks and achieve the ideal balance of parallel performance and optimized availability, Harvard University’s Faculty of Arts and Sciences Research Computing (FASRC) deployed the DataDirect Networks (DDN®) GRIDScaler® GS7KX® parallel file system appliance with 1PB of storage. The installation has sped the collection of images detailing synaptic connectivity in the brain’s cerebral cortex.
News from insideBIGDATA
- Machine Learning Managed Services: Can Big Tech Provide Viable Alternatives to IIoT Predictive Maintenance Software?
- Parallel Storage Fuels Groundbreaking Neuroscience and Behavioral Research at Harvard
- Deep Threat Analysis: Going Beyond the Obvious to Improve Border Security
- Accelerate Business Value with Serverless Computing
- DDN and Parabricks Announce Solution for GPU Acceleration of Genome Analysis
- Atomic X Unveils ‘Rufus’, An Intelligent AI-Powered Chatbot
- ArrayFire Releases v3.6 Announced
Editor’s Choice
Celebrating 20 Years of the OpenMP API
"The first version of the OpenMP application programming interface (API) was published in October 1997. In the 20 years since then, the OpenMP API and the slightly older MPI have become the two stable programming models that high-performance parallel codes rely on. MPI handles the message passing aspects and allows code to scale out to significant numbers of nodes, while the OpenMP API allows programmers to write portable code to exploit the multiple cores and accelerators in modern machines." [Read More...]
Quantum Computing: Its Principles, Capabilities and Challenges
Dr. Mark Mattingley-Scott from IBM gave this talk at the Swiss HPC Conference. "Quantum Computing is here, right now - and we are at the start of a new way of computing, which will impact us the way the revolution started by Shockley, Bardeen and Brattain did in 1947. In this talk I will introduce Quantum Computing, its principles, capabilities and challenges and provide you with the insight you need to decide how you should engage with this revolutionary technology." [Read More...]
New AI Performance Milestones with NVIDIA Volta GPU Tensor Cores
Over at the NVIDIA blog, Loyd Case shares some recent advancements that deliver dramatic performance gains on GPUs to the AI community. "We have achieved record-setting ResNet-50 performance for a single chip and single server with these improvements. Recently, fast.ai also announced their record-setting performance on a single cloud instance. A single V100 Tensor Core GPU achieves 1,075 images/second when training ResNet-50, a 4x performance increase compared to the previous generation Pascal GPU." [Read More...]
NVIDIA Simplifies Building Containers for HPC Applications
In this video, CJ Newburn from NVIDIA describes how users can benefit from running their workloads in the NVIDIA GPU Cloud. "A container essentially creates a self contained environment. Your application lives in that container along with everything the application depends on, so the whole bundle is self contained. NVIDIA is now offering a script as part of an open source project called HPC Container Maker, or HPCCM that makes it easy for developers to select the ingredients they want to go into a container, to provide those ingredients in an optimized way using best-known recipes." [Read More...]
Announcing the Dell EMC AI Challenge
In this video from the Dell EMC HPC Community meeting, Shreya Shar and Onur Celebioglu from Dell describe the new Dell EMC AI Challenge. "Dell EMC invites you to participate in an AI Challenge! Abstracts and proposals will be accepted through May 31, 2018, and will be reviewed according to the following three criteria: Most innovative use of AI technology, Biggest economic or social impact, and Highest potential for commercialization." [Read More...]
