Materials Innovation at Argonne

 New looks at innovation with AI and HPC at Argonne.

Automatic Database Creation for Materials Discovery: Innovation From Frustration

Argonne National Laboratory  by John Spizzirri

Scientists at the U.S. Department of Energy's Argonne National Laboratory and the U.K.'s University of Cambridge have developed a method that generates automatic databases to support specific scientific disciplines, using artificial intelligence and high-performance computing (HPC). The technique can assemble databases via natural language processing (NLP) and HPC, most of which was performed at the Argonne Leadership Computing Facility. The team built a database on both material structures and material properties, using the NLP ChemDataExtractor data-mining application. Cambridge's Jacqueline Cole said, "It's probably the first such compilation of a database on such a massive scale, with 5,380 like-for-like pairs of experimental and calculated data. And because it's such a large amount, it serves as a repository in its own right and really opens the door to predicting new materials."

Quantum Accelerated Computation as Disruption

Some bits of thoughts on the topic.   And new form of high performance computing (HPC)?  Ultimately what are the limits of capabilities here, versus more standard supercomputing?

Is Quantum-Accelerated Computation the Next Big Disruption?  by 7wData

Many people are looking to quantum computing as the next revolutionary technology. Nature analyzed that in 2017 and 2018 alone, more than $450 million of private funding was poured into the quantum industry. Even the classical finance community starts to smell an opportunity. Xavier Rolet, the former CEO of the London Stock Exchange and well-respected industry veteran, told The Quantum Daily that he considers such investments a solid bet on the future and believes in the transformational change of quantum computers.

If not all, the exciting topic made its way to a more mainstream audience. Even the tabloids have been writing extensively and with very catchy headlines about a Nature article published in 2019. Researchers at Google announced that they achieved what is called quantum supremacy. On their quantum processor named Sycamore (see Fig. 1), they ran some calculations within 200 seconds that would have taken the world’s most powerful (classical) supercomputer 10,000 years — at least they claim. It has to be added that the setup was very specific and the results are heavily debated by competitor IBM. But certainly, the expectation towards the field has been starting to skyrocket.

As smart and quirky physicists move towards the field of quantum computation, build hyped startups and get huge funding, it is very interesting to follow this space. Will we have the chance to see disruptive innovation live and in action? ... (More with sign in to 7wData ) ... '

Simulating Mask Effectiveness

Was alway surprised this was not better understood.  Had for years done simulation of complex systems using high performance computing methods.    Why not for mask applications?    Too many possible variants in initial conditions?   Here what seems to be a something related, but with supercomputing.  Can this kind of simulation could be improved with quantum computing?

Do Cloth Masks Work? Supercomputer Fugaku Says Yes
Nikkei Asian Review
Yuki Misumi

Japan's Riken Institute said the Fugaku supercomputer, recently crowned the world's fastest, developed a model that showed nonwoven fabric masks block virus-laden respiratory droplets more effectively than cotton or polyester masks (although all three types were deemed effective at slowing the spread of the coronavirus). The system simulated the performance of the three types of fabric masks in blocking the spray of virus-carrying respiratory droplets from coughing by the wearer, demonstrating that all three types stopped at least about 80% of spray. The team also simulated a virus spreading through a 2,000-seat auditorium, and found little danger of proliferation if visitors are masked and sitting spaced apart.... ' 

AI and Aerospace

Model based systems engineering is along time approach in this space. we did much of i in the enterprise.  AI/machine learning can be used to focus subtasks with broader process or resource implications. 

Modeling and simulation: Achieve next-level results with AI
Artificial intelligence-based approaches are key to realizing model-based systems engineering benefits.

Aerospace executives can now optimize manufacturing processes by leveraging artificial intelligence (AI) with high-performance computing (HPC) technologies and the digital thread. A digital thread follows the lifecycle of a product from design inception through engineering and product lifecycle management, to manufacturing instructions, supply chain management, and through to service events. You'll be able to enhance the aerospace design process to protect budgets, avoid static production rates, and nudge your business ahead of competitors.

Even better, as aerospace design becomes more complex, AI can help keep your business ahead of the innovation curve.... '

Quantum Simulation from D-Wave

Been watching D-Wave closely through advances and challenges, since their inception,   here another advance, in particular linking to material science.

D-Wave Demonstrates First Large-Scale Quantum Simulation of Topological State of Matter    From Inside HPC

D-Wave Systems researchers have demonstrated a topological phase transition using a 2048-quantum bit (qubit) annealing quantum computer, a complex quantum simulation of materials that marks a major step toward reducing the need for time-consuming physical research and development. The team says the techniques used in this work could have far-reaching implications in the development of novel materials. D-Wave's Mohammad Amin says this breakthrough represents the first theoretically predicted state of matter to be realized in quantum simulation before being demonstrated in a real magnetic material.  ... 

High Performance Codes

As the volume and complexity increase, and as we try to improve more complex processes we need more high performance computing (HPC).  Its a complex thing to build code that we know will operate correctly and predict well at very large scale.

'A System Purely for Developing High-Performance, Big Data Codes' 
Rice News by Jade Boyd

Rice University researchers this week will debut the PlinyCompute platform at the 2018 ACM SIGMOD conference in Houston, TX. Developed under Rice's Pliny Project banner, the PlinyCompute tool was created to ease implementation of complex objects and workflows on big data platforms. Rice's Jia Zou says PlinyCompute differs from the Spark platform "because it was designed for high performance from the ground up." Compared to Spark, benchmarking showed PlinyCompute was at least twice as fast, and in some cases 50 times faster, at implementing complex object manipulation and library-style computations, Zou says. The Pliny project is aimed at creating programming tools that can "autocomplete" and "autocorrect" code.  ... "

IBM Advances PowerAI Enterprise Solution

IBM's New PowerAI Features Again Demonstrate Enterprise AI Leadership  by Patrick Moorhead  in Forbes.

What appears to be a semi automation of various areas of machine learning, to make them easier to create and maintain and work with relevant data. The claim is with minimal ML expertise.  Notable to me is one emphasis on the analysis of image and video data.

" ... Recently, IBM announced a significant revamp of PowerAI, seeking to address some of the bigger challenges facing developers and data scientists—cutting down the time required for AI system training significantly, and simplifying the development experience. .... " 

IBM writes about this, and describes the system, which requires POWER HPC hardware, you can test it free at the link:

" .. PowerAI makes deep learning, machine learning, and AI more accessible and more performant. By combining this software platform for deep learning with IBM® Power Systems™, enterprises can rapidly deploy a fully optimized and supported platform for machine learning with blazing performance. The PowerAI platform includes the most popular machine learning frameworks and their dependencies, and it is built for easy and rapid deployment. PowerAI requires installation on IBM Power Systems S822LC for HPC server infrastructure ... " 

See also their blog on OpenPower, and a description of advances.

A Look at last Years HPC Conference

In the CACM:   I used to cover High Performance Computing (HPC), topics for the enterprise.   Now the topic has seemed to be overtaken by volume rather than performance.  But no,  Torsten Hoefler overviews the recent conference.

Higher Performance Computing and a New Paradigm

Via Doug McDavid,

Big data spawns a new scientific paradigm.  From Pacific Northwest Labs, which we worked with on text analytics visualization.

Presentation

High Performance Computing for Data Intensive Science by John R. Johnson
Computational Sciences & Mathematics Division
Fundamental & Computational Sciences Directorate

High Performance Computing in CPG

Video by a former colleague,  Tom Lange, about how High Performance Computing (HPC) is used in consumer packaged goods at Procter & Gamble.  " ....     From toothpaste to diapers to dish soap, companies like Procter & Gamble use high performance computing to create better, more efficient, more cost effective products and services to improve our everyday lives. ... "  And more on HPC in other domains. 

Dell Diversifies the Supply Chain

I have not looked at Dell Computer for some time, was reintroduced to some of their high performance computing offerings.  And here, more about their supply chain.   " ... As a business diversifies, it may need more than one supply chain. Here’s how Dell addressed that challenge. .. " 

Visualizing Data from High Performance Computing

Simulation creates a great deal of data, so it naturally links well with visualization   Every industrial simulation we wrote for the enterprise was linked to a visualization to understand the subtleties of parameter changes and experimentation.    In Computer Org:   Largely non technical coverage.

" ... For computational simulations, the era of "big data" ended before it began. We're actually living in the era of infinite data — in which the stream pouring forth from computational models and simulations can be as voluminous as outputing every value at every timestep, drowning disks and researchers alike with high-cadence, arbitrarily large datasets. Rather than struggling to make models bigger, the challenge is now to keep them under control.  ... " 

Links to a number of other articles and resources.  See link below for more coverage in this blog on HPC.

High Performance Computing, Everywhere

In the CACM:  High performance computing (HPC)  An overview.  From a November conference. HPC is in broad use today, we all gets its benefits everyday.  " ... If you drive a car, take medicine, use shampoo,  fly in airplanes, watch movies, expect your washer to arrive at your door undented, have a loved one coping with Alzheimer's Disease, use a cellphone,  care about the structure and fate of the universe, or worry about providing clean drinking water for the inhabitants of the planet, then your quality of life is directly affected — improved — by HPC. ... ".  

The Wikipedia has a good overview of the topic.  We actively used the method in the enterprise.

Big Data and Big Science

In CACM:  About the integration of high performance computing (HPC) and Big Data.  As the article suggests,   the result of this is to create more data.   And I add,  more opportunity for analytics.

Related Technical Paper. 
Synergistic Challenges in Data-Intensive Science and Exascale Computing
Summary Report of the Advanced Scientific
Computing Advisory Committee (ASCAC), Subcommittee  March 30, 2013

High Performance Computing in the Cloud

Featured in CN Computing Now:    " ... High performance computing is no longer limited to those who own supercomputers. HPC's democratization has been driven particularly by cloud computing, which has given scientists access to supercomputing-like features at the cost of a few dollars per hour. The four articles I've selected for this month's Computing Now theme highlight some current work in academia and industry, examining key benefits, challenges, and enablers of HPC in the cloud.... " 

High Performance Computing Clusters

Was just introduced to HPCC Systems.  As an open source solution to a 'Big Data' problem to support a large scale enterprise simulation.  Well worth a look for those interested in the topic.  More to follow.