Research Computing - MGHPCC Feed

July Publications

Below is a selection of papers that appeared in July 2019 reporting the results of research using the Massachusetts Green High Performance Computing Center (MGHPCC), or acknowledging the use of Harvard’s Odyssey Cluster, Northeastern’s Discovery Cluster, the Boston University Shared Computing Cluster and MIT’s Engaging Cluster all of which are housed at the MGHPCC.

Anders Andreassen, Ilya Feige, Christopher Frye, and Matthew D. Schwartz (2019), Binary JUNIPR: an interpretable probabilistic model for discrimination, arXiv: 1906.10137 [hep-ph]

T.T.G. Beucler (2019), Interaction between water vapor, radiation and convection in the tropics, MIT Doctoral Dissertation

Blakesley Burkhart and Philip Mocz (2019), The Self-gravitating Gas Fraction and the Critical Density for Star Formation, The Astrophysical Journal, doi: 10.3847/1538-4357/ab25ed

Phillip A. Cargile, Charlie Conroy, Benjamin D. Johnson, Yuan-sen Ting, Ana Bonaca, Aaron Dotter (2019), MINESweeper: Spectrophotometric Modeling of Stars in the Gaia Era, arXiv: 1907:07690 [astro-ph.SR]

Stephen Carr, Chenyuan Li, Ziyan Zhu, Efthimios Kaxiras, Subir Sachdev, and Alex Kruchkov (2019), Coexistence of ultraheavy and ultrarelativistic Dirac quasiparticles in sandwiched trilayer graphene, arXiv:1907.00952 [cond-mat.str-el]

Cheng-Chieh Chuang, Hsun-Chen Chu, Sheng-Bor Huang, Wei-Shun Chang, Hsing-Yu Tuan (2019), Laser-induced plasmonic heating in copper nanowire fabric as a photothermal catalytic reactor, Chemical Engineering Journal, doi: 10.1016/j.cej.2019.122285

Charlie Conroy, Ana Bonaca, Phillip Cargile, Benjamin D. Johnson, Nelson Caldwell, Rohan P. Naidu, Dennis Zaritsky, Daniel Fabricant, Sean Moran, Jaehyon Rhee, Andrew Szentgyorgyi (2019), Mapping the Stellar Halo with the H3 Spectroscopic Survey, arXiv: 1907.07684 [astro-ph.GA]

Qian Di, Heresh Amini, Liuhua Shi, Itai Kloog, Rachel Silvern, James Kelly, M. Benjamin Sabath, Christine Choirat, Petros Koutrakis, Alexei Lyapustin, Yujie Wang, Loretta J. Mickley, Joel Schwartz (2019), An ensemble-based model of PM2.5 concentration across the contiguous United States with high spatiotemporal resolution, Environmental International, doi: 10.1016/j.envint.2019.104909

Richard M. Feder, Stephen K. N. Portillo, Tansu Daylan, and Douglas Finkbeiner (2019), Multiband Probabilistic Cataloging: A Joint Fitting Approach to Point Source Detection and Deblending, arXiv: 1907.04929 [astro-ph.IM]

William Fitzhugh, Fan Wu, Luhan Ye, Haoqing Su, Xin Li (2019), Strain‐Stabilized Ceramic‐Sulfide Electrolytes, Small, doi: 10.1002/smll.201901470

Joel Leja, Sandro Tacchella, and Charlie Conroy (2019), Beyond UVJ: More efficient selection of quiescent galaxies with UV / mid-IR fluxes, arXiv: 1907.02970 [astro-ph.GA]

Xiaorong Liu, Jianhan Chen (2019), Residual Structures and Transient Long-Range Interactions of p53 Transactivation Domain: Assessment of Explicit Solvent Protein Force Fields, Journal of Chemical Theory and Computing, doi: 10.1021/acs.jctc.9b00397

Andreea Panaitescu, Meng Xin, Benny Davidovitch, Julien Chopin, and Arshad Kudrolli (2019), Birth and decay of tensional wrinkles in hyperelastic sheets, arXiv: 1906.10054 [cond-mat.soft]

Joonha Park, Yves Atchad (2019), Markov chain Monte Carlo algorithms with sequential proposals, arXiv: 1907.06544 [stat.CO]

Sun, Qinfang (2019), Simulating Hydrogen Bonded Clusters and Zeolite Clusters for Renewable Energy Applications, University of Massachusetts, Amherst, Doctoral Dissertation.

Sherer, Zachary (2019), A Comparison of Two Architectures for Breadth-First Search on FPGA SoC Platforms, University of Massachusetts, Lowell, Masters Dissertation

Tirado, Luis Eladio (2019), On-the-move Detection of Security Threats Using 3D MM-Wave Radar Imaging, Northeastern University Doctoral Dissertation

Do you have news about research using computing resources at the MGHPCC? If you have an interesting project that you want to tell people about or a paper you would like listed, contact hlh@mit.edu

Links

MGHPCC Publications

HolyokeCodes: Soccer Robots

Students at a HolyokeCodes Soccer Robotics program learn to play soccer with robots using JavaScript.

RoboCup is an international scientific initiative with the goal of advancing the state of the art of intelligent robots. When established in 1997, the original mission was to field a team of robots capable of winning against the human soccer World Cup champions by 2050.

Of the multiple Robocup Soccer leagues that exist, the “Small Size” league is one of the oldest. Arjun Guha and Joydeep Biswas, assistant professors in the College of Information and Computer Sciences at UMass Amherst, and their students adapted the UMass Minute Bots Robocup Team robots (which they use to compete in Small Size league competition), to a JavaScript interface designed for high school students. Centrally controlled via radio, using perception based on a central overhead camera, the robots can travel up to 5 m/s and kick, chip-kick, and dribble a golf ball.

Over the course of a week in early July, local area students in grades 9 through 12 learned basic commands to control the robots, developed simple planning algorithms, and programmed behaviors for offensive and defensive roles with the week culminating in a series of 2 v 2 matches. Students came away from the activity with a new appreciation for the problems of intelligent multi-robot/agent cooperation and control in the highly dynamic environment of even a toy soccer pitch plus plenty of hands-on coding and robotics experience.

Links

Holyoke Codes

Container Gardening

On June 27, 2019, Holyoke Community College celebrated the graduation of 12 apprentices from its Freight Farms workforce training program, a new apprenticeship-based opportunity to learn the art and practice of cutting-edge hydroponics based in a facility close to the MGHPCC.

Training takes place in a pair of refurbished shipping containers  “Freight Farms” located on Race Street, in the heart of Holyoke. The re-purposed containers are used to grow leafy greens and herbs without the use of soil. Each of the container farms can hold 256 grow towers with a capacity of 10-12 plants each, and are able to grow as much produce in a year as an acre of farmland.

The soil-free facilities use water, mineral nutrients and LED lights for growing leafy greens like lettuce – image courtesy Holyoke Community College.

From learning how to seed, transplant, harvest and package crops to maintaining a safe, clean, and organized work area, apprentices graduate the program with a full working knowledge of how to work in a hydroponic facility and follow food safety standards

The Holyoke Freight Farm is a collaboration between Holyoke Community College, Nuestras Raices and the City of Holyoke, with support from MassDevelopment’s Transformative Development Initiative. MGHPCC is happy to be a neighbor and to assist, as the fiscal sponsor for a TDI Fellows Cohort from MassDevelopment.

Northeast CyberTeam to Host Two “Birds of a Feather” Sessions at PEARC19

Chicago, July 29, 2019 – The Northeast CyberTeam Initiative, a National Science Foundation-funded project to expand access to high performance computing among researchers at small and mid-sized colleges and universities, will host two “Birds of a Feather” sessions at PEARC19 beginning today.

MGHPCC Press Release

The first session, which will take place Monday from 5:15 p.m. to 6:15 p.m., will focus on ASK.CI, developed by the Cyberteam in partnership with the XSEDE Campus Champions to aggregate answers to a broad spectrum of questions that are commonly asked by the research computing community. The Platform has participation from nearly 300 volunteers representing over 240 institutions worldwide and has attracted over 100,000 page views since launch. The PEARC session will include a discussion of lessons learned since launch and plans for the coming year, in particular the introduction of “locales”, institution-specific categories created under the umbrella of Ask.CI.

The second session, scheduled for Tuesday from 5:15 p.m. to 6:15 p.m., will provide information for audience members interested in learning about the Cyberteam Initiative. Recognizing that Research Computing Facilitators (RCFs) are often critical to enabling the successful transition to advanced computing, and that RCFs are common at large universities and corporations, but rare at smaller institutions, the Cyberteam Initiative aims to fill the gap by giving interested students hands-on RCF experience, creating a pipeline of talent to meet growing academic and industry demand for that role. The Cyberteam has launched 28 projects over the past two years that pair student RCFs with experienced mentors to work with researchers and educators. In parallel, the project has developed a portal, that provides access to web-based self-service learning resources such as Ask.CI, as well as tools to manage projects. The portal encapsulates knowledge and experience that has been gained, with the goal of making it possible to replicate the methodology in other regions.

“The Northeast Cyberteam Initiative has developed a robust methodology for making research computing more readily accessible to smaller institutions in the region” said John Goodhue, Principal Investigator of the Northeast Cyberteam Initiative and Executive Director of the Massachusetts Green High Performance Computing Center. “Our goal is to find opportunities to replicate or adapt the methodology for use in other parts of the country.”

“While Ask.CI has shown steady growth in both contributions and audience, we believe unmet need for it still exists,” said Julie Ma, Program Manager of both the Northeast Cyberteam Initiative and Ask.CI. “We strive to continue to raise awareness of ASK.CI to meet that need while expanding on the depth and breadth of the content.”

About PEARC
The Practice Experience in Advanced Research Computing (PEARC) Conference Series is a community-driven effort built on successes of the past, with the aim to grow and be more inclusive by involving additional local, regional, national, and international cyberinfrastructure and research computing partners spanning academia, government and industry. The PEARC conference series is working to integrate and meet the collective interests of our growing community by providing a forum for discussing challenges, opportunities and solutions among the broad range of participants in the research computing community. The PEARC conferences are organized by a group of dedicated volunteers from the community.

About the Northeast Cyberteam Initiative
High performance computing has become an in indispensable part of scientific inquiry today. The Northeast Cyberteam Initiative seeks to build a sustainable system of computational support for researchers at small and mid-sized colleges and universities in New England, who typically lack the computing resources available at larger institutions. The Cyberteam Initiative offers online computing tools and is developing a regional pool of Research Computing Facilitators (RCFs), who are expert at connecting researchers with appropriate computer systems. Funded by the National Science Foundation, the Cyberteam is a collaborative project led by the Massachusetts Green High Performance Computing Center, University of Maine System, University of New Hampshire, and University of Vermont, with support from Worcester Polytechnic Institute.

About the XSEDE Campus Champions Program
The Campus Champions are a community of practice of campus research computing professionals celebrating 10 years since initiation under Teragrid and support through XSEDE. The Campus Champions have grown to over 470 individuals from over 250 institutions, including 46 Minority Serving Institutions and 70 institutions in Established Program to Stimulate Competitive Research (EPSCoR) jurisdictions. The Campus Champions focus discussion on challenges, opportunities, solutions, and leading practices via an active mailing list and videoconferences. The Campus Champions community promotes and facilitates the effective participation of a diverse national community of campuses in the application of advanced digital resources and services to accelerate scientific discovery and scholarly achievement.

About the Massachusetts Green High Performance Computing Center
The Massachusetts Green High Performance Computing Center (MGHPCC) provides state-of-the-art infrastructure for computationally intensive research that is indispensable in the increasingly sensor and data-rich environments of modern science and engineering. Computers at the MGHPCC run millions of virtual experiments every month, supporting thousands of researchers in Massachusetts and around the world. The MGHPCC was developed through an unprecedented collaboration among the most research-intensive universities in Massachusetts (Boston University, Harvard University, the Massachusetts Institute of Technology, Northeastern University and the University of Massachusetts); the Commonwealth of Massachusetts; and private industry (Cisco and Dell EMC). The member universities fund the ongoing operation of the data center, which is open for use by any research organization.

###

Media Contact:
Jennifer Rosenberg
781-854-2997
jenn@howellcomm.com

Related

Have a Research Computing Question? Just “Ask”
MGHPCC Press Release

Northeast Cyberteam Initiative Announces Measures to Help Researchers at Small and Mid-Size N.E. Colleges and Universities Leverage All-Important High Performance Computing Resources
MGHPCC Press Release

Edging Towards a Greener Future

Materials scientists at Harvard use computers at the MGHPCC to design better solid-state lithium ion batteries through advanced characterizations and simulation.

As electric cars have evolved, running costs have begun to approach those for their gasoline forebears – more expensive to buy, cheaper to run – but range and charge time remain an outstanding issue. Leading the list of possible game-changers are solid-state lithium ion batteries in which wet electrolyte is replaced with a solid. Simpler, such units have the potential to be a lot cheaper, lighter, and not require liquid cooling. They are expected to also be longer-lasting and fireproof while potentially being much faster-charging too.

A major impediment to the development of solid‐state lithium‐ion batteries, however, are interfacial reactions between ceramic‐sulfide solid‐electrolytes and common electrodes. Such reactivity means that ceramic‐sulfide batteries require a suitable coating material to isolate the electrolyte from the electrode materials.

Xin Li, a materials scientist at Harvard, is focused on the design of new energy-related materials through advanced characterizations and simulations. By combining electrochemistry, electron microscopy, X-ray diffraction, and first-principles simulations, his team uses Harvard’s Odyssey Cluster (housed at the MGHPCC in Holyoke, MA) to understand the relationship between atomistic structure and electrochemical property of materials.

In new work published in the journal Advanced Energy Materials Li, working with graduate student William Fitzhugh and others in Harvard’s Paulson School of Engineering, computationally evaluated the interfacial stability of the lithium sulfide Li10SiP2S12 with over 67 000 potential coating materials. Their study found 2000 materials predicted to form stable interfaces in the cathode voltage range and over 1000 materials for the anode range. Studies like these are invaluable in narrowing the field of potential candidate materials ahead of physical trials.

Systematic assessment through first-principles analysis of large systems like this is of course computationally highly demanding. Innovations introduced in the paper included a new binary‐search algorithm to improve the speed and accuracy for evaluating pseudo and the authors highlight the computational challenges posed by high‐throughput interfacial phase‐diagram calculations as well as pragmatic computational methods they would recommend to make such calculations routinely feasible.

In addition to the over 3000 materials cataloged, representative materials from the anionic classes of oxides, fluorides, and sulfides were also chosen to experimentally demonstrate chemical stability when in contact with Li10SiP2S12. “We chose LiCoO2 as an example cathode material to identify coating compounds that would be stable with both Li10SiP2S12 and a common cathode,” Li explains. “Analyzing the correlation between elemental composition and multiple chemical and electrochemical instability metrics revealed key trends in, amongst others, the role of anion selection.”

Xin says the next step in this work will focus on computationally determining and experimentally realizing the best interface coating material by considering more advanced effects in all-solid-state batteries.

To find out more about this work contact Xin

About the Researcher Associate Professor of Materials Science
Xin Li

Xin Li is Assistant Professor of Materials Science in the John A. Paulson School of Engineering and Applied Sciences at Harvard University. His research group focuses on the design of new energy-related materials through advanced characterizations and simulations. By combining electrochemistry, (in situ) electron microscopy, (in situ) X-ray diffraction and first-principles simulations, his team seeks to understand the relationship between atomistic structure and electrochemical property of materials.

Publication

William Fitzhugh, Fan Wu, Luhan Ye, Wenye Den, Pengfei Qi, Xin Li (2019), A High‐Throughput Search for Functionally Stable Interfaces in Sulfide Solid‐State Lithium Ion Conductors, Advanced Energy Materials, doi: 10.1002/aenm.201900807

Related

The Culprit of Superconductivity in Cuprates, John A. Paulson School of Engineering and Applied Science News, Harvard University

Links

Xin Li

Li Laboratory

Convergent regulatory evolution & loss of flight in paleognathous birds

Study using Harvard’s Odyssey Cluster housed at the MGHPCC explores the genetics behind how paleognathous birds lost the ability to fly.

Read this story in the Harvard Gazette

Since Darwin’s era, scientists have wondered how flightless birds like emus, ostriches, kiwis, cassowaries, and others are related, and for decades the assumption was that they must all share a common ancestor who abandoned the skies for a more grounded life.

By the early 2000s, new research using genetic tools upended that story, and instead pointed to the idea that flightlessness evolved many times throughout history. Left unanswered, however, were questions about whether evolution had pulled similar or different genetic levers in each of those independent avian lineages.

A team of Harvard researchers believes they may now have part of the answer.

Based on an analysis of the genomes of more than a dozen flightless birds, including an extinct moa, a team of researchers led by Tim Sackton, director of bioinformatics for the FAS Informatics Group, and Professor of Organismic and Evolutionary Biology Scott Edwards found that while different species show wide variety in the protein-coding portions of their genomes, they appear to turn to the same regulatory pathways when evolving flight loss. The study is described in a recent paper published in Science.

In addition to Sackton and Edwards, the study was co-authored by Jun Liu, professor of statistics and of biostatistics; statistics research assistant Zhirui Hu; Alison Cloutier, a postdoctoral researcher working in Edwards’ lab; Phil Grayson, a graduate student in the Edwards lab; and teams from New Zealand, the University of Texas, Austin, and the Royal Ontario Museum.

“There is a long history in evolutionary biology of converging traits — the idea that there’s independent evolution toward the same kind of phenotype,” Sackton said. “What we were interested in is, how does that happen?”

Flightless birds all have similar body types, Sackton noted. “They have reduced forelimbs [wings], to different degrees, and they all have this loss of the ‘keel’ in their breastbone that anchors flight muscles,” he said. “What that amounts to is a suite of convergent morphological changes that led to this similar body plan across all these species.”

To understand what drove that suite of changes, Sackton, Edwards, and their colleagues turned to the genomes of the birds themselves.

“We wanted to compare not just the parts of the genome that code for proteins, but also the parts of the genome that regulate when those proteins are expressed,” Sackton said. To identify those regions in the various species examined for the study, the team used a process that involved aligning the genomes of more than three dozen bird species — both flying and flightless — and then identifying regions that showed relatively few differences in their genetic sequence. These places in the genome that are conserved but are not part of proteins, are likely to have a regulatory function.

“We worked with collaborators in Statistics here at Harvard to develop a new statistical method that allowed us to ask, for each of those regulatory elements, how many of these species showed the same pattern of divergence, suggesting they have changed the same regulatory elements,” Sackton said. “And what we found was that while there is not much sharing of protein-coding genes [in general], there is for these regulatory regions, suggesting that there are shared developmental pathways that are repeatedly targeted every time this phenotype has evolved.”

While the protein-coding genes appear to be responsible for adaptations in diet, feather function, and environment, Sackton said, the regulatory regions seem to play a key role in the body-scaling changes that go along with flight loss.

“What’s interesting about the morphological changes … is they have to preserve their hind limbs,” he said. “There are lots of ways to stop a limb from forming, but shrinking a forelimb without changing the hind limb is more difficult.”

And in some ways, Sackton said, that makes sense — strange as it may seem, it is likely easier to simply not form a limb rather than shrink one.

“If you think about it, there’s lots of ways to break something,” he said. “There are a bunch of steps early in limb development where, if a protein doesn’t get expressed, it’ll just turn the system off and you don’t get a limb. But this is actually a complicated shift in body scaling. You can’t just willy-nilly grow limbs to different sizes, so … the fact that it’s important they maintain functional hind limbs constrains the system and might be why we see this convergent pattern.”

To prove that theory, the team tagged certain regulatory regions in the birds’ genomes with a gene that would produce green fluorescent protein. They found that in flightless species, where those regions were believed to have undergone functional changes, the marker gene was effectively turned off.

“To get a limb to start growing, a bunch of things need to happen … so if you can knock out an enhancer and make it harder for those proteins to be expressed you can delay that process,” Sackton said. “This suggests these regions may have lost some important binding sites that prevent them from acting as an enhancer.”

What it boils down to, Sackton said, is that birds have a limited number of options to pursue when it comes to the loss of flight, and so various species have gone to the same well again and again.

“That’s the conclusion we would draw from this work,” he said. “There are a limited number of ways you can get this type of change in scaling, and they center on this regulation of early limb development.”

The study also highlights the power of the multidisciplinary approach taken by Sackton, Edwards, and their colleagues.

“One of the things that was exciting about this project for me personally was how we were able to bring the computational expertise in the Informatics Group to bear on this really important question in evolutionary biology,” said Sackton. “This joining of computational, statistical genetics with the natural history perspectives is important for getting the full picture of how these birds evolved.”

“It’s exciting what can be done with a research team with diverse skill sets,” Edwards added. “Our group had developmental biologists, computational biologists, morphologists, statisticians, population geneticists — and, of course, ornithologists. Each brings a different perspective and the results, I think, are amazing.”

This research was supported with funding from the National Science Foundation and the Natural Sciences and Engineering Research Council of Canada.

Story image: Zhirui Hu (from left), Scott Edwards, Tim Sackton, and Philip Grayson discuss their new study – credit: Kris Snibbe/Harvard Staff Photographer

Publication

Timothy B. Sackton, Phil Grayson, Alison Cloutier, Zhirui Hu, Jun S. Liu, Nicole E. Wheeler, Paul P. Gardner, Julia A. Clarke, Allan J. Baker, Michele Clamp, Scott V. Edwards (2019), Convergent regulatory evolution and loss of flight in paleognathous birdsScience, doi: 10.1126/science.aat7244

 

How to Capture Waste Heat Energy with Improved Polymers

UMass Amherst chemists, electrical engineers use computing facilities at the MGHPCC to identify new variable for material design.

Read this story at UMass Amherst News

AMHERST, Mass. – By one official estimate, American manufacturing, transportation, residential and commercial consumers use only about 40 percent of the energy they draw on, wasting 60 percent. Very often, this wasted energy escapes as heat, or thermal energy, from inefficient technology that fails to harvest that potential power.

Now a team at the University of Massachusetts Amherst led by chemist Dhandapani Venkataraman, “DV,” and electrical engineer Zlatan Aksamija, report this month in Nature Communications on an advance they outline toward more efficient, cheaper, polymer-based harvest of heat energy.

“It will be a surprise to the field,” DV predicts, “it gives us another key variable we can alter to improve the thermo-electric efficiency of polymers. This should make us, and others, look at polymer thermo-electrics in a new light.”

Aksamija explains, “Using polymers to convert thermal energy to electricity by harvesting waste heat has seen an uptick in interest in recent years. Waste heat represents both a problem but also a resource; the more heat your process wastes, the less efficient it is.” Harvesting waste heat is less difficult when there is a local, high-temperature gradient source to work with, he adds, such as a high-grade heat source like a power plant.

Thermo-electric polymers are less efficient at heat harvesting compared to rigid, expensive-to-produce inorganic methods that are nevertheless quite efficient, Aksamija adds, but polymers are worth pursuing because they are cheaper to produce and can be coated on flexible materials – to wrap around a power plant’s exhaust stack, for example.

Recently, scientists have been addressing this obstacle with a process called “doping.” With it, researchers mix chemical or other components into polymers to improve their ability to move electric charges and boost efficiency. DV says, “Imagine that we’ve added chocolate chips, a material that improves conductivity, to a cookie. That’s doping.”

But doping involves a tradeoff, Aksamija adds. It can either achieve more current and less thermally-induced voltage, or more voltage and less current, but not both. “If you improve one property, you make the other worse,” he explains, “and it can take a lot of effort to decide the best balance,” or optimal doping.

To address this, DV and his chemistry Ph.D. student Connor Boyle, with Aksamija and his electrical engineering Ph.D. student Meenakshi Upadhyaya worked in what DV calls “a true collaboration,” where each insight from numerical simulations informed the next series of experiments, and vice versa.

The chemists conducted experiments, while the engineering team performed efficiency analyses along the curve from “zero doping” to “maximum doping” to identify the best balance for many different materials. For the massive number of simulations they ran to test hundreds of scenarios, they used the Massachusetts Green High Performance Computing Center in nearby Holyoke.

UMass team of chemists and electrical engineers outline a new way to advance a more efficient, cheaper, polymer-based harvest of heat energy to produce electricity. Graphic courtesy of UMass Amherst/Meenakshi Upadhyaya.

Aksamija says. “We can now tell you, for every given material, what is the optimal balance of the two properties, and for a while, people were satisfied with just knowing that.” But along the way, he adds, they discovered an entirely new variable that had not yet been accounted for, one that turned out to be critical to the doped polymer’s ability to harvest thermal energy efficiently.

He says, “The original analysis didn’t get at the question of the position of the doping components, whether the materials clump or not and how much they clump, or cluster, as we call it. It turns out that clustering is a critical variable.” The team turned to chemist Michael Barnes, a co-author on their recent paper, who used Kelvin Probe Force Microscopy to probe the dopants at the nano level and show that clustering is indeed present in polymers doped at room temperature, but not at higher temperatures.

With that confirmation, the researchers turned to modeling an expanded trade-off curve, says Upadhyaya. From their theoretical modeling, she and Aksamija found that clustering alters the shape of that curve. To improve efficiency beyond the current-voltage trade-off, one must move the entire trade-off curve, she says.

This unexpected finding should provide a new path for designing more efficient polymers for thermo-electric devices, the researchers say. DV notes that until now, chemists and materials scientists have been trying to organize polymers to be more like the inorganics, “nicely aligned and very regular, which is difficult to do,” he adds. “It turns out that this may not be the way to go; you can take another road or another approach. We hope this paper provides a basis to move polymer-based thermo-electrics forward.”

Publication

Connor J. Boyle et al. (2019), Tuning charge transport dynamics via clustering of doping in organic semiconductor thin films, Nature Communications, doi: 10.1038/s41467-019-10567-5

Related

A Little Bit of This… A Little Bit of That… MGHPCC News

Links

The DV Group

June Publications

Below is a selection of papers that appeared in June 2019 reporting the results of research using the Massachusetts Green High Performance Computing Center (MGHPCC), or acknowledging the use of Harvard’s Odyssey Cluster, Northeastern’s Discovery Cluster, and the Boston University Shared Computing Cluster all of which are housed at the MGHPCC.

Sylvain Capponi et al (2019), NMR relaxation in the spin-1 Heisenberg chain, arXiv: 1905.12697 [cond-mat.str-el]

A. Devarakonda, H. Inoue, S. Fang,  C. Ozsoy-Keskinbora, T. Suzuki, M. Kriener, L. Fu, E. Kaxiras, D. C. Bell, and J. G. Checkelsky (2019), Evidence for clean 2D superconductivity and field-induced finite-momentum pairing in a bulk vdW superlattice, arXiv:19809.02065 [cond-mat.supr-con]

Mingu Kang et al. (2019), Dirac fermions and flat bands in the ideal kagome metal FeSn, arXiv:1906.02167

A. Mesgarnejad, A. Karma (2019), Phase Field Modeling of Chemomechanical Fracture of Intercalation Electrodes: Role of Charging Rate and Dimensionality, arXiv:1906.07655 [cond-mat.mtrl-sci]

Pavitra Muralidhar (2019), Mating preferences of selfish sex chromosomes, Nature, doi: 10.1038/s41586-019-1271-7

Shilo Ohayon, Arik Girsault, Maisa Nasser, Shai Shen-Orr, Amit Meller (2019), Simulation of single-protein nanopore sensing shows feasibility for whole-proteome identification, PLOS Computational Biology, doi: 10.1371/journal.pcbi.1007067

Andres N. Salcedo, Benjamin D. Wibking, David H. Weinberg, Hao-Yi Wu, Douglas Ferrer, Daniel Eisenstein, and Philip Pinto (2019), Cosmology with Stacked Cluster Weak Lensing and Cluster-Galaxy Cross-Correlations, arXiv:1906.06499 [astro-ph.CO]

Hui Shao, Wenan Guo and Anders W. Sandvik (2019), Monte Carlo Renormalization Flows in the Space of Relevant and Irrelevant Operators: Application to Three-Dimensional Clock Models, arXiv: 1905.13640 [cond-mat.str-el]

Joshua Stern, Qingqing Xiong, Anthony Skjellum, and Martin C. Herbordt (2019), A Novel Approach to Supporting Communicators for In-Switch Processing of MPI Collectives

Su, Jiaying (2019), A Numerical Study of Heat Transfer in a Square Channel With and Without Bleed Holes Roughened With Staggered Square Ribs, Masters Thesis, Northeastern University

Mirac Suzgun, Sebastian Gehrmann, Yonatan Belinkov, Stuart M. Shieber (2019), LSTM Networks Can Perform Dynamic Counting, arXiv:1906.03648 [cs.CL]

Waleed Tahir, Ulugbek S. Kamilov, and Lei Tian (2019), Holographic particle localization under multiple scattering, Advanced Photonics, doi: 10.1117/1.AP.1.3.036003

Isabelle R Weir, Lu Tian, Ludovic Trinquart (2019), Multivariate meta-analysis model for the difference in restricted mean survival times, Biostatistics, doi: 10.1093/biostatistics/kxz018

Qingqing Xiong, Rushi Patel, Chen Yang, Tong Geng, Anthony Skjellum, Martin C. Herbordt (2019), GhostSZ: A Transparent FPGA-Accelerated Lossy Compression Framework

Yifan Zhang, Lorenzo Trippa, Giovanni Parmigiani (2019), Frequentist operating characteristics of Bayesian optimal designs via simulation, Statistics in Medicine, doi: 10.1002/sim.8279

Do you have news about research using computing resources at the MGHPCC? If you have an interesting project that you want to tell people about or a paper you would like listed, contact hlh@mit.edu

Links

MGHPCC Publications

Physics-driven Drug Discovery

Reporting by Helen Hill for MGHPCC News

MGHPCC industry partner Silicon Therapeutics’ innovative research uses the Center’s facilities to host its physics-based simulation platform for advanced drug discovery and design.

Silicon Therapeutics, a fully integrated physics-driven drug discovery company, is one of several Massachusetts-based biotechnology companies now partnering with the Massachusetts Green High Performance Computing Center (MGHPCC) to capitalize on the Center’s pre-eminent research computing facilities.

Simulating biological systems at the atomic scale reveals important information about drug targets. Silicon Therapeutics is using their proprietary physics-based simulation platform to advance therapeutics toward curative outcomes for patients with currently unmet medical need, with a particular focus on innate immunity in cancer and inflammation. Their initial drug discovery program has produced first-in-class small molecule STING agonists that shows activity across all known isoforms of the human STING protein and demonstrate potent anti-tumor activity in vivo.

“At Silicon, we are running massive simulations based on quantum mechanics and molecular dynamics to design novel molecules for previously undrugged targets,” explains Dr Woody Sherman, the company’s CSO. “We also run simulations to determine the effect of disease-driving genetic mutations on protein conformation and how that relates to functional activity, which allows us to devise our drug design strategy for key biological targets.”

Biological molecules such as proteins are constantly in motion in a complex environment consisting of water, cofactors (chemicals that assist enzymes during the catalysis of reactions), ions, and other biological molecules. In many cases, the dynamic nature of proteins is essential for their function. Sherman says, “The physics-based simulation platform we have developed closely represents the real dynamic nature of protein targets, enabling us to gain deep insights into the biomolecular recognition process associated with targets that were previously considered ‘undruggable’.”

Silicon Therapeutics’ current high-performance computing (HPC) platform is comprised of both internal and cloud resources, however,  most of the company’s drug discovery workloads run on Neo, a cluster with more than 300 GPUs and over a thousand CPUs housed in five racks at MGHPCC. Neo’s first-generation nodes have Intel Xeon processors for a total of 16 cores and 64 GB of memory. The more recent nodes have Intel Skylake processors, with 16 cores and 96 GB of memory. Altogether, the system has almost 3 petaflops of peak single-precision performance from almost a million Nvidia GPU cores for running CUDA and OpenCL jobs. Neo runs Ubuntu OS and utilizes Slurm for job scheduling. “It has been a pleasure to work with MGHPCC,” says Sherman. “It is a state-of-the-art research computing facility run by HPC veterans who know the challenges associated with executing big workloads. In addition, housing dedicated computing resources at the MGHPCC has allowed us to scale our compute infrastructure to the level needed to maximally impact our drug discovery projects.”

The first drug discovery project at Silicon Therapeutics has focused on STING (Stimulator of Interferon Genes), a master regulator of type I interferons and a key mediator of innate immunity. Activation of STING provides two critical anti-tumor features:

  1. The “spark” for initiating a robust innate immune response.
  2. Enhancement of the adaptive immune response.

Using its platform, Silicon Therapeutics has successfully designed first-in-class small molecule STING agonists that show activity across all known isoforms of the human STING protein and demonstrate potent anti-tumor activity in vivo. Using their physics-driven discovery engine, plus cutting-edge capabilities in biology, chemistry, and biophysics, Silicon Therapeutics has developed small molecule agonists of STING that exhibit activity across all known isoforms of the human protein. When dosed via intravenous administration our small molecule STING agonists show potent anti-tumor activity in mice bearing syngeneic tumors.

Other drug discovery projects in the Silicon Therapeutics pipeline can be found on their website: https://silicontx.com/pipeline/

Links

https://silicontx.com/

Related

Three Massachusetts Biotech Companies Leverage MGHPCC

 

 

 

 

NEREN Seminar: “Bridging the Gap: Sharing Computing Resources Across Campuses”

The Northeast Research and Education Network, NEREN’s, April 5 seminar held at the Providence Marriott.

The seminar, the sixth in a series of day-long seminars devoted to proposing and advancing ideas for regional collaboration in research computing and networking kicked off with an update from Barr von Oehsen, Associate Vice President of the Office of Advanced Research Computing (OARC) at Rutgers University on progress building ERN (Eastern Regional Network), a collaboration of regional research and educational networks and their members from the eastern portion of the US seeking to federate high performance computing resources.

The second presentation, given by Abraham Matta, Ph.D., Professor and Chair of Computer Science, Boston University, introduced the GENI (Global Environment for Networking Innovation) Project a virtual laboratory spanning compute resources distributed across the United States.

Thomas A. DeFanti, Ph.D., Research Scientist at UCSD’s Calit2/Qualcomm Institute and Professor Emeritus of Computer Science, University of Illinois then reflected on how lessons learned through building community computing infrastructure on the Pacific Research Platform, might translate to the context of a national platform.

Finally Paul Earsy, Senior Sales Engineer, Confluent shared a deep dive into what makes Apache Kafka® a popular tool for real-time data collecting, delivery, and processing.

A breakout session concluded the meeting.

Presentations were recorded and are available at the webcast link below.

AGENDA | WEBCAST (requires registration)

NEREN is a consortium of non-profit organizations that provide a fiber-optic network connecting and unifying the research and education communities in New York and New England.

Typically NEREN sponsors a twice-yearly seminar intended to foster such collaboration.

Industry sponsors included Intel, Red RiverDell, Mainline Information Systems, Juniper Networks, and Cohesity.

To learn about other events, or if you have questions, please email Laurie Robinson, NEREN Program Administrator, at laurie@neren.org or by phone: 401-523-5107.

Image credit: G. Silva

Related

NEREN Seminar: “Bridging the Gap: Advancing Regional Collaboration and Research IT Collaboration” MGHPCC News

 

Mayor Announces Winners of MGHPCC Scholarships

Total Awards Top $50k, Part of Broader Collaboration on Education & Workforce Development.

MGHPCC press release 

Holyoke, Massachusetts, June 6, 2019 – Holyoke Mayor Alex B. Morse today announced the winners of two scholarship awards totaling $5,000 from the Massachusetts Green High Performance Computing Center (MGHPCC) to Holyoke Public School students. Now in its sixth year, the MGHPCC merit-based scholarship program has awarded more than $50,000 in grants so far.
“I’m grateful to the MGHPCC and its member institutions for their support of the scholarship program, which is just one part of our collaboration around education and workforce training,” said Morse. “We’re fortunate to have such great partners.”

“Providing educational and training opportunities to local students is an important part of our mission as well as our commitment to Holyoke,” said John Goodhue, executive director of the MGHPCC. “The scholarship program is particularly rewarding because it helps smooth the path for deserving students who have demonstrated a desire to make the extra effort to succeed.”

This year’s scholarship recipients are Helena Middleton and Sarah Sirkissoon, both will attend UMass Amherst in the fall. Helena will major in biology and Sarah will study economics and political science.

“Sarah and Helena are hard-working students who earned this recognition and support, and I have every confidence that they will succeed in their academic careers and beyond,” said Dr. Stephen Zrike, superintendent of schools. “We are grateful to the MGHPCC for creating this opportunity for deserving students.”

The MGHPCC scholarship program is open to any Holyoke Public School student enrolling at one of the five founding institutions of the MGHPCC: Boston University, Harvard University, Northeastern University, MIT or UMass. It is open to students in any major and not limited to students of computer science or other scientific disciplines. Recipients are chosen based on an 800-word essay answering the question, “What do you see as the three biggest challenges currently facing Holyoke, and what would you do to address them?”

In addition to sponsoring the scholarship program, the MGHPCC works with the mayor’s office, Holyoke Public Schools, Holyoke Community College, Springfield Technical Community College and local community service organizations to host internships, robotics competitions, teacher workshops, after-school classes, programming courses, career awareness days and summer camps.

Related

MGHPCC Awards $10K in Scholarships to Local Graduating Seniors MGHPCC News

$16,000 in Scholarships Given to 4 Holyoke High School Seniors by Computing Center MGHPCC News

 

May Publications

New Publications Last Month

Below is a selection of papers that appeared in May 2019 reporting the results of research using the Massachusetts Green High Performance Computing Center (MGHPCC), or acknowledging the use of Harvard’s Odyssey Cluster, Northeastern’s Discovery Cluster, and the Boston University Shared Computing Cluster all of which are housed at the MGHPCC.

Asanga Bandara, Afra Panahi, George A. Pantelopulos, Tetsuro Nagai, and John E. Straub (2019), Exploring the impact of proteins on the line tension of a phase-separating ternary lipid mixture, The Journal of Chemical Physics, doi: 10.1063/1.5091450

Ligia R. Benavides, Julia G. Cosgrove, Mark S. Harvey, Gonzalo Giribet (2019), Phylogenomic interrogation resolves the backbone of the Pseudoscorpiones Tree of Life, Molecular Phylogenetics and Evolution, doi: 10.1016/j.ympev.2019.05.023

Pak‐Wah Chan, Pedram Hassanzadeh, Zhiming Kuang (2019), Evaluating Indices of Blocking Anticyclones in Terms of Their Linear Relations With Surface Hot Extremes, Geophysical Research Letters, doi: 10.1029/2019GL083307

Sarah C. Conner, Lisa M. Sullivan, Emelia J. Benjamin, Michael P. LaValley, Sandro Galea, Ludovic Trinquart (2019), Adjusted restricted mean survival times in observational studies, Statistics in Medicine, doi: 10.1002/sim.8206

Feng Ding and Robin D. Wordsworth (2019), A New Line-by-line General Circulation Model for Simulations of Diverse Planetary Atmospheres: Initial Validation and Application to Exoplanet GJ 1132B, arXiv: 1905.04635 [astro-ph.EP]

William Fitzhugh, Fan Wu, Luhan Ye, Wenye Deng, Pengfei Qi, Xin Li (2019), A High‐Throughput Search for Functionally Stable Interfaces in Sulfide Solid‐State Lithium Ion Conductors, Advanced Energy Materials, doi: 10.1002/aenm.201900807

Barak Gabai, Xi Yin (2019), On The S-Matrix of Ising Field Theory in Two Dimensions, arXiv: 1905.00710 [hep-th]

Brandt A.L. Gaches, Stella S.R. Offner, Thomas G. Bisbas (2019), The Astrochemical Impact of Cosmic Rays in Protoclusters I: Molecular Cloud Chemistry, arXiv: 1905.02232 [astro-ph.GA]

Gregory M. Green, Edward Schlafly, Catherine Zucker, Joshua S Speagle, Douglas Finkbeiner (2019), A 3D Dust Map Based on Gaia, Pan-STARRS 1 and 2MASS, arXiv: 1905.02734 [astro-ph.GA]

Jonathan I. Leckenby, Miranda A. Chacon, Adriaan O. Grobbelaar, Jeff W. Lichtman (2019), Imaging Peripheral Nerve Regeneration: A New Technique for 3D Visualization of Axonal Behavior, Journal of Surgical Research, doi: 10.1016/j.jss.2019.04.046

Xinzhi Li, Amit Das, Dapeng Bi (2019), Mechanical heterogeneity in tissues promotes rigidity and controls cellular invasion, arXiv: 1905.02697 [physics.bio-ph]

Austin Thai and Sheryl M. Grace (2019), Prediction of small quadrotor blade induced noise, 25th AIAA/CEAS Aeroacoustics Conference, 20-23 May, 2019, Delft, The Netherlands, doi: 10.2514/6.2019-2684

Thomas L. Williams, Steven A. Lopez, Leila F. Deravi (2019), A Sustainable Route To Synthesize the Xanthommatin Biochrome via an Electro-catalyzed Oxidation of Tryptophan Metabolites, ACS Sustainable Chem. Eng., doi: 10.1021/acssuschemeng.9b01144

Xiao Wu, Yi Xu, Bradley P. Carlin (2019), Optimizing Interim Analysis Timing for Bayesian Adaptive Commensurate Designs, arXiv: 1905.07456 [stat.AP]

Phelan Yu, Lawrence W. Cheuk, Ivan Kozyryev, and John M. Doyle (2019), A Scalable Quantum Computing Platform Using Symmetric-Top Molecules, arXiv: 1905.06439 [physics.atom-ph]

Krissia Zawadzki and Adrian E. Feiguin (2019), Time and momentum-resolved tunneling spectroscopy of pump-driven non-thermal excitations in Mott insulators, arXiv: 1905.08166 [cond-mat.str-el]

Jiawei Zhuang, Daniel J. Jacob, Judit Flo Gaya, Robert M. Yantosca, Elizabeth W. Lundgren, Melissa P. Sulprizio, Sebastian D. Eastham (2019), Enabling immediate access to Earth science models through cloud computing: application to the GEOS-Chem model, Bulletin of the American Meteorological Society, doi: 10.1175/BAMS-D-18-0243.1

Do you have news about research using computing resources at the MGHPCC? If you have an interesting project that you want to tell people about or a paper you would like listed, contact hlh@mit.edu

Links

MGHPCC Publications

UMass Lowell Hosts 5th Annual HPC Day

The HPC Day conference provides a showcase of computational research in science, engineering, and computer science research being carried out in the New England area. This full-day conference which took place this year on May 21, 2019 at UMass Lowell featured speakers from leading research and education institutes across the region. 

This year’s keynote speaker was Kirk Jordan (IBM) with a presentation entitled “Data Centric Systems: Algorithm Exploitation & Evolving AI/Cognitive Examples”.

Speaking the morning session was Mark Hempstead (Tufts University) “Workload Characterization Tools for Every Need: From Architecture Agnostic Classification of Communication to Trace-based Simulation of Multi-Threaded Workloads”, Stephen de Bruyn Kops (UMass Amherst) “Why huge simulations are invaluable for understanding fluid flow physics”, and Mary Jo Ondrechen (Northeastern University) “Electrostatic networks in natural enzymes: What can we learn for protein engineering?”

After a lunch sponsored by Dell EMC, after speakers included Julia Levites (NVIDIA) “GPU Hackathons – accelerating applications through hands-on experience and collaboration”, Stratis Ioannidis (Northeastern University) “Distributing Frank-Wolfe via Map-Reduce”, Nurit Haspel (UMass Boston) “Detecting Large Scale Chromosomal Rearrangements: A big data challenge”, Noah Van Dam (UMass Lowell) “Using HPC for engine and fuel spray simulations”, Maricris Mayes (UMass Dartmouth) “Quantum chemical study of the initial-assembly of aromatic dipeptides into nanosturectures for biomedical applications”, Benoit Forget (MIT) “High-fidelity nuclear reactor simulations and the need for Exascale computing”, and Dmitry Korkin (Worcester Polytechnic Institute) “Frozen: Finding genomic elements that are extremely conserved in evolution using cache-oblivious computing”.

This year the prize for best-poster presentation went to

Djeneba Kassambara (University of Massachusetts Lowell) “Accurate Numerical Solutions of Helmholtz Equations Using Layered Media Green’s Function.”

The conference was hosted by the Mathematical Sciences Department at UMass Lowell in collaboration with and support from  the Massachusetts Green High Performance Computing Center (MGHPCC), and several industry partners including NVIDIA Dell EMC, and SIAM.

Story image credit: the Ondrechen Lab, Northeastern University

Related Links

Northeastern University Hosts 4th Annual HPC Day MGHPCC News

UMass Dartmouth Hosts 3rd Annual HPC Day MGHPCC News

UMass Dartmouth Hosts the Second Statewide HPC Day MGHPCC News

 

 

 

April Publications

New Publications Last Month

Below is a selection of papers that appeared in April 2019 reporting the results of research using the Massachusetts Green High Performance Computing Center (MGHPCC), or acknowledging the use of Harvard’s Odyssey Cluster, Northeastern’s Discovery Cluster, and the Boston University Shared Computing Cluster all of which are housed at the MGHPCC.

B. A. Cook, Charlie Conroy, Pieter van Dokkum, and Joshua S. Speagle (2019), Measuring Star-formation Histories, Distances, and Metallicities with Pixel Color-Magnitude Diagrams I: Model Definition and Mock Tests, arXiv:1904.00011 [astro-Ph.GA]

Jing Guo, Guangyu Sun, Bowen Zhao, Ling Wang, Wenshan Hong, Vladimir A. Sidorov, Nvsen Ma, Qi Wu, Shiliang Li, Zi Yang Meng, Anders W. Sandvik, and Liling Sun(2019), Quantum phases of SrCu2(BO3)2 from high-pressure thermodynamics, arXiv:1904.09927 [cond-mat.str-el]

Paul Michael Kielstra and Marius Lemm (2019), On the Finite-size Lyapunov Exponent for the Schrodinger Operator with Skew-shift Potential, arXiv: 1904:08871 [math-ph]

Marius Lemm, Anders Sandvik, and Sibin Yang (2019), The AKLT model on a hexagonal chain is gapped, arXiv:1904.01043 [quant-ph]

Mariana Levi, Jeffrey K. Noel, Paul C. Whitford (2019), Studying ribosome dynamics with simplified models, Methods, doi: 10.1016/j.ymeth.2019.03.023

Ziqiang Lü, Haiying Gao, Jianshe Lei, Xiaotao Yang, Sampath Rathnayaka, Cong Li (2019), Crustal and upper mantle structure of the Tien Shan orogenic belt from full‐wave ambient noise tomography, Journal of Geophysical Research – Solid Earth, doi: 10.1029/2019JB017387

Dimitrios Maroudas and Brian D Wirth (2019), Atomic-scale modeling toward enabling models of surface nanostructure formation in plasma-facing materials, Current Opinion in Chemical Engineering, doi: 10.1016/j.coche.2019.03.001

Julian B. Mu˜noz (2019), A Standard Ruler at Cosmic Dawn, arXiv:1904:07868 [astro-ph.CO]

Julian B. Mu˜noz (2019), Velocity-induced Acoustic Oscillations at Cosmic Dawn, arXiv:1904:07881 [astro-ph.CO]

Deena A. Rennerfeldt, Joana S. Raminhos, Samantha M. Leff, Pristinavae Manning, Krystyn J. Van Vliet (2019), Emergent heterogeneity in putative mesenchymal stem cell colonies: Single-cell time-lapsed analysis, PLOS One, doi: 10.1371/journal.pone.0213452

Timothy B. Sackton, Phil Grayson, Alison Cloutier, Zhirui Hu, Jun S. Liu, Nicole E. Wheeler, Paul P. Gardner, Julia A. Clarke, Allan J. Baker, Michele Clamp, Scott V. Edwards (2019), Convergent regulatory evolution and loss of flight in paleognathous birds, Science, doi: 10.1126/science.aat7244

Waleed Tahir, Jiabei Zhu, Sreekanth Kura, Xiaojun Cheng, David Boas, and Lei Tian (2019), A Deep Learning Approach to 3D Segmentation of Brain Vasculature, in Biophotonics Congress: Optics in the Life Sciences Congress 2019 (BODA,BRAIN,NTM,OMA,OMP), The Optical Society (Optical Society of America, 2019), paper BT2A.6.

Bowen Zhao, Phillip Weinberg and Anders W. Sandvik (2019), Symmetry-enhanced discontinuous phase transition in a two-dimensional quantum magnet, Nature Physics, doi: s41567-019-0484-x

Do you have news about research using computing resources at the MGHPCC? If you have an interesting project that you want to tell people about or a paper you would like listed, contact hlh@mit.edu

Links

MGHPCC Publications

Three Massachusetts Biotech Companies Leverage MGHPCC

Partnerships Support Trend of Increased Collaboration and Knowledge Transfer between Industry and Public High Performance Computing Centers

Holyoke, Massachusetts, April 29, 2019 – Reflecting a national trend of increased collaboration between industry and high performance computing centers, three Massachusetts-based biotechnology companies have engaged the Massachusetts Green High Performance Computing Center (MGHPCC) to enable their leading-edge work, the MGHPCC announced today.

“Science and engineering today rely on high performance computing. Our mission is to support scientific and engineering advances as well as the state’s innovation economy by enabling computationally intensive research across academia, government and the private sector,” said John Goodhue, the Executive Director of the MGHPCC.

The companies using the MGHPCC include Boston-based Silicon Therapeutics, the first fully integrated physics-driven drug discovery company; IOMICS, an innovative life science analytics company headquartered in Cambridge; and Curoverse, a Somerville-based unit of Veritas Genetics dedicated to building open source technology to accelerate health care’s transformation to a more predictive personalized and precise model.

“Computing is central to our physics-driven drug discovery efforts at Silicon Therapeutics,” said Vipin Sachdeva, Associate Director and Head of High Performance Computing at Silicon Therapeutics. “Since our inception, MGHPCC has proved to be pivotal in addressing the computing needs of our growing organization. They have demonstrated excellence across the board in our interactions with them and we are excited to continue to expand our HPC infrastructure at the MGHPCC.”

“The MGHPCC has emerged as one of the Commonwealth’s most important resources for the 21st Century,” said Joseph Gormley, Chief Technology Officer, and Senior Software Architect at IOMICS. “The center has been a catalyst for multiple commercial-academic collaborations and has been instrumental in the development of our award-winning life science analytics products and services.”

In addition to the MGHPCC’s partnerships with companies such as Silicon Therapeutics, IOMICS and Curoverse, the founding universities of the MHGPCC collaborate with a wide range of commercial enterprises through joint initiatives such as the Center for Data Science at UMass Amherst, which is creating new technology to manage and gain insight from big data while also educating tomorrow’s data scientists, and the Mass Open Cloud, a partnership led by Boston University to create a new model of cloud computing that allows academic researchers and technology companies to innovate in ways that are not currently possible with commercial clouds.

These partnerships reflect a national trend of increased collaboration and scientific knowledge transfer between industrial users and public high performance computing centers, according to a study funded by the National Science Foundation (NSF) and conducted by Hyperion Research (formerly the IDC HPC Group) for the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign.

“This study represents the increased awareness and importance of HPC centers and industry collaborations, and the recognition that these partnerships need to be better understood,” said Bill Kramer, Senior Associate Director, Blue Waters Project Office of the NCSA. “One of the most interesting observations to me is that the industrial partners and centers both value the HPC expertise that centers provide to the industrial project as much as the actual computing/analysis system access.”

 

About the Massachusetts Green High Performance Computing Center
The Massachusetts Green High Performance Computing Center (MGHPCC) provides state-of-the-art infrastructure for computationally intensive research that is indispensable in the increasingly sensor and data-rich environments of modern science and engineering. Computers at the MGHPCC run millions of virtual experiments every month, supporting thousands of researchers in Massachusetts and around the world. The MGHPCC was developed through an unprecedented collaboration among the most research-intensive universities in Massachusetts (Boston University, Harvard University, the Massachusetts Institute of Technology, Northeastern University and the University of Massachusetts); the Commonwealth of Massachusetts; and private industry (Cisco and Dell EMC). The member universities fund the ongoing operation of the data center, which is open for use by any research organization.

###

Media Contact:
Jennifer Rosenberg
781-854-2997
jenn@howellcomm.com

 

Modeling Plasma-Surface Interactions

by Helen Hill for MGHPCC News

Designing plasma-facing components (PFCs) that can tolerate the extreme heat and particle flux exposure conditions inside a fusion reactor core is one of the major obstacles toward the practical realization of nuclear fusion. Dimitrios Maroudas, a Professor in the Chemical Engineering Department at the University of Massachusetts Amherst, uses MGHPCC computing facilities to study plasma-surface interactions and their effects on surface morphology and near-surface structure in plasma-facing components of nuclear fusion reactors.

Dimitrios Maroudas is a Professor in the Chemical Engineering Department at the University of Massachusetts Amherst. His interests lie in the area of multi-scale modeling of complex systems with special emphasis on theoretical and computational materials science and engineering.

Professor Dimitrios Maroudas is the Director of the Materials Engineering Program in the Department of Chemical Engineering at UMass Amherst.

“We use high-performance computing to simulate materials processing and function,” explains Maroudas. “Our work seeks to predict the structure, properties, and reliability of electronic and structural materials with the goal of informing optimal design of nanostructured materials and metamaterials for applications in electronic technologies and various energy technologies.”

In addition to obtaining a fundamental understanding of the behavior of complex material systems, his team is especially interested in modeling processing and function of thin-film, nanostructured, and low-dimensional forms of materials used in device fabrication technologies. “All of these material systems are characterized by structural inhomogeneities, such as crystalline lattice imperfections, surfaces, interfaces, and a variety of nanostructural features,” says Maroudas. “Understanding the formation and evolution of such nano/micro-structure and morphology during physical or chemical processing and during device function is particularly important in developing processes that yield optimal material properties and guarantee device performance and reliability.”

Research efforts in the Maroudas Group focus on the development and implementation of computational quantum, statistical, and continuum mechanical methods for the study of structure and dynamics and for predictions of bulk and interfacial properties of heterogeneous materials. “In my group we place special emphasis on establishing rigorous links between atomistic and macroscopic (continuum) length scales and between fast and slow time scales,” says Maroudas. “This allows us to develop coarse-grained descriptions of multi-scale, multi-physics phenomena in complex materials starting from an atomistic, first-principles-based description of bonding and dynamics.”

Unsurprisingly, research in his group employs a broad range of computational methods from electronic structure calculation techniques to continuum numerical modeling, including: ab-initio calculations of atomic structure, total energy, and atomic-scale dynamics based on density functional theory; structural relaxation, lattice-dynamics, Monte Carlo, and molecular-dynamics simulation methods based on empirical and semi-empirical descriptions of interatomic interactions; kinetic Monte Carlo and mean-field rate equation models; and continuum modeling techniques based on domain discretization using finite-element, finite-difference, boundary-element, and spectral methods. In addition, his group employs analytical and numerical stability and bifurcation theory as a means of understanding materials’ structural and morphological response to variations in processing and operating parameters. “A special interest on this front is the development of methods for overcoming time-scale limitations of atomistic dynamical simulators and enabling such simulators to perform numerical bifurcation and stability analysis,” says Maroudas.

A current research topic is the study of plasma-surface interactions and their effects on surface morphology and near-surface structure in plasma-facing components of nuclear fusion reactors.

Designing plasma-facing components (PFCs) that can tolerate the extreme heat and particle flux exposure conditions inside a fusion reactor core is one of the major obstacles toward the practical realization of nuclear fusion.

In a recent article published in Current Opinion in Chemical Engineering, Maroudas used computing facilities housed at the Massachusetts Green High Performance Computing Center (MGHPCC) to perform the atomic-scale simulations they review in their paper to provide a fundamental understanding of the dynamical response of tungsten, an important PFC material, to reactor-relevant plasma exposure conditions leading to helium implantation. The paper’s co-author is Brian Wirth, a Professor in the Department of Nuclear Engineering at the University of Tennessee, Knoxville, and lead PI of the DOE-funded SciDAC (Scientific Discovery through Advanced Computing) center on Plasma-Surface Interactions that supports this research effort.

The hierarchy of simulations Maroudas and Wirth report included molecular-statics computations to establish helium-surface interaction energetics and the origin of helium segregation on tungsten surfaces, targeted molecular-dynamics (MD) simulations of near-surface helium cluster reactions, and large-scale MD simulations of implanted helium evolution in plasma-exposed tungsten conducted in leadership-scale computing facilities.

The results of their studies reveal that cluster-surface elastic interactions induce drift fluxes of small mobile helium clusters in the tungsten toward the plasma-exposed tungsten surface, which facilitate helium segregation on the surface and activate cluster reactions, most importantly trap mutation, which generates a flux of self-interstitial tungsten atoms to the surface. Near-surface cluster dynamics of this type has a significant effect on PFC surface morphological evolution, near-surface defect structure formation, especially the nucleation and growth of helium nanobubbles, and the amount of helium retained in the PFC material upon plasma exposure.

“Such a mechanistic understanding enables the development of atomistically-informed coarse-grained models of surface nanostructure formation in PFC materials,” Maroudas says. “A crucial step toward predicting PFC surface degradation and improving PFC operating lifetime and reactor performance.”

Story image: Schematic of the International Thermonuclear Experimental Reactor, ITER – image courtesy: D. Maroudas

Publication:

Dimitrios Maroudas and Brian D Wirth (2019), Atomic-scale modeling toward enabling models of surface nanostructure formation in plasma-facing materials, Current Opinion in Chemical Engineering, doi: 10.1016/j.coche.2019.03.001

Links:

Dimitrios Maroudas

Dimitri Maroudas is the UMass PI in multi-million, multi-institutional SciDAC Center on Plasma-Surface Interactions – UMass Amherst, Chemical Engineering Department Newsletter, Spring 2018

 

The MGHPCC Supercloud

By Helen Hill for MGHPCC News

The MGHPCC Supercloud, operated by MIT Lincoln Laboratory Supercomputing Center (LLSC) is a facility intended to be of particular use to people who need to transition from their desktop workstation to a high performance computing cluster in order to solve larger or more complex problems, but do not have the time to gain in-depth specialized knowledge about parallelizing application software, running batch jobs, optimizing startup times for large collections of simultaneous jobs, and other details that end users must otherwise master in a Linux cluster environment.

Managed by MIT, but as a separate enterprise dealing with security-related (classified) research, Lincoln Laboratory began as a center for radar and early warning system research and has a long history of advanced technologies in sensors (signals intelligence, image recognition, etc) and associated computing. Today the Laboratory (now the largest US Department of Defense federally funded research and development center) is home to numerous unclassified open research activities, for example, development of the imaging devices and lenses for TESS or LIDAR-based tools now providing FEMA with automated post-disaster operational damage analysis.

The MGHPCC Supercloud project grew out of a desire to strengthen collaboration between Lincoln Laboratory and regional partners in the MGHPCC Consortium (BU, Harvard, MIT, Northeastern, and UMass). LLSC having already demonstrated its success in providing a similar cloud service to Lincoln Laboratory researchers, opened the doors to provide its service to members of the MGHPCC consortium.

The Supercloud supports most of the desktop computing environments that are commonly used in the research and engineering community today, including Matlab, R Studio, Jupyter Notebooks, and Machine Learning frameworks such as Tensorflow, Caffe, and Theano.

“If there’s something a user needs and we don’t already have it or something equivalent to it we will work with the user to help them get set up with what they need,” says Lauren Milechin, who is in charge of new-user liaison.

In particular, the MGHPCC Supercloud provides a software platform which allows users to launch large-scale interactive compute jobs from their desktop, as well as facilitating ready sharing of large volumes of project data. The Supercloud experience enables reference datasets to be pre-positioned in databases, while also providing access to software modules and training to reduce user ramp-up time, all within a responsive, interactive supercomputing environment.

Simple line commands allow users to tailor hardware selection (ie from a single processor core to a whole node), using LLMapReduce (an implementation of native MapReduce), carry out language agnostic parallel data analysis or run Parallel Matlab. Users are also able to couple Supercloud resources with web-based interactive development environments like the popular Jupyter Notebook, or manage dynamic databases from a simple GUI.

“While you can run traditional big MPI Fortran code on the MGHPCC Supercloud, our really big success has been our ability to get really large numbers of people with little if any prior parallel computing background effectively working on these systems in very short order,” says Jeremy Kepner, a Lincoln Laboratory Fellow and Founding Head of the LLSC.

High-performance data analysis (HPDA) requires high-level programming environments but also rapid interaction and fast turnaround. By optimizing every aspect of its HPDA system, the LLSC TX-Green system running on 32 000 cores (512 x 64-core Xeon nodes) with the same setup and software stack as the MGHPCC Supercloud has been demonstrated to be able to launch 260 000+ data analytics in just 40 seconds ie 6000+ launches per second, or a 1000x speed-up over standard approaches.

In another measure, some analytic applications are written uniquely for Microsoft Windows. Standard approaches based on Virtual Machines or Windows HPC can take hours to launch across thousands of cores. MGHPCC Supercloud launch system optimizations enable launch of 16 000 + Microsoft Windows environments (running WINE) on 16 000+ cores (256 x 64-core Xeon nodes) in just five minutes: ie 50+ launches per second or a 100x speed-up over standard approaches.

Request Your Account

visit supercloud.mghpcc.org/requesting-account where a member of the Supercloud team is standing by to help you migrate your application to their system. Links

MGHPCC Supercloud

Lincoln Laboratory Supercomputing Center

Viewpoint: High-performance computing center is asset worth recognizing

President of  Boston University and Chair of the Massachusetts Green High Performance Computing Center writing today in the Boston Business Journal.

By Robert Brown

Robert Brown is the president of Boston University and chair of the Massachusetts Green High Performance Computing Center.

Massachusetts is known the world over as a hub of scientific discovery and innovation. Each year, the commonwealth’s research enterprise attracts billions of dollars in funding, creating thousands of jobs while spinning out new products and companies. Less known, even within Massachusetts, is the role and importance of research computing — the massive computational infrastructure on which this research relies.

Research computing has become as essential to knowledge discovery as theory or real-world experiment. Today, virtually no breakthrough or advance — from health care to geophysics to urban planning — takes place without trillions of computations. Research computing is now a mainstay of the state’s innovation economy, as important as any single field or industry.

A decade ago, recognizing the importance of high-performance computing to both science and the economy, an unprecedented partnership among academia, state and local government, and the technology sector launched a world-class data center called the Massachusetts Green High Performance Computing Center (MGHPCC). In partnership with the commonwealth, the majority funding for construction and ongoing operations has come from the five founding universities — Boston University, Harvard University, MIT, Northeastern University, and the University of Massachusetts system, along with essential support from industry partners Cisco and EMC, and the federal government.

As I begin my final year as chair of the MGHPCC board of directors, it is my hope that the value of research computing and the MGHPCC continue to be recognized and leveraged by policymakers and the business community as a key asset and partner in the state’s efforts to attract public and private research funding (for which competition is intense) and the economic benefits it brings.

The MGHPCC is used by tens of thousands of researchers who perform millions of virtual experiments every month in every field of scientific endeavor. Some examples:

BU assistant professor Jonathan Appavoo is working with Boston Children’s Hospital radiologist Ellen Grant on a project to deliver high-quality fetal scans to prospective parents in seconds rather than hours.

A team led by Harvard professor Brendan Meade is using neural networks to better predict earthquake aftershock locations and better understand the physics that cause them.

MIT professor Adam Willard is modeling the movement of electrons through organic photovoltaic material in work that could change the future of solar energy.

Northeastern assistant professor Paul Whitford is modeling complex proteins to improve our understanding of biological processes and to enable new nanotechnology applications.

UMass Amherst chemist Scott Auerbach is working on the development of renewable, carbon-neutral fuels from biomass.

In addition to creating national-scale computational capacity, the MGHPCC provides a platform for its members to integrate computer science research, operations expertise, and industry partnerships to advance the state of the art in research computing services and support. The MGHPCC has also fostered regional collaborations such as the Northeast Cyberteam initiative, an NSF-funded partnership between the MGHPCC and other universities in the region to improve support for researchers at small and mid-sized institutions.

Located in Holyoke, the MGHPCC has become an important partner in education and workforce development with public schools in the Pioneer Valley. Programs include internships for local community college students and after-school activities that introduce middle school students to computing concepts.

Research computing is an often-unheralded service performed by an unseen network of hardware, software, and people, but it is core to the modern research infrastructure that undergirds the steady stream of scientific breakthroughs coming out of our laboratories and to the economic transformation of Massachusetts.

Computer Science Teachers Association Western Mass Chapter

The Computer Science Teachers Association Western Mass chapter hosted a free workshop for teachers on Microsoft MakeCode for Minecraft on December 15 at the Massachusetts Green High Performance Computing Center.  The education version of Minecraft lets students participate in and build an open-world game to promote creativity, collaboration, problem-solving and coding skills.  “Microsoft MakeCode for Minecraft” is a block programming platform specially designed for teachers and students to make things happen in the game Minecraft (for example raining chickens!). The workshop, led by John Nguyen, Microsoft Learning Consultant, introduced twelve local Computer Science teachers to this platform.  “Using a hands-on approach to computing education, Microsoft MakeCode brings computer science to life for all students with fun projects, immediate results, and both block and text editors for learners at different levels.”

The education version of Minecraft lets students participate in and build an open-world game to promote creativity, collaboration, problem-solving and coding skills.

The Computer Science Teachers Association (CSTA) is a membership organization that supports and promotes the teaching of computer science. CSTA’s mission is to empower, engage and advocate for K-12 CS teachers worldwide.  Membership consists of more than 25,000 members from more than 145 countries.

The Western Massachusetts chapter of the CSTA is run by CS teachers throughout the region and organizes workshops and events to support computer science teachers.  In addition to this workshop, this year CSTA-WMass also presented a Scratch Meetup, a workshop on teaching with micro:bit, a panel discussion on College Computing Programs in the Age of CS for All, and helped teachers prepare for CS Education Week in December.

A  Code.org CS Fundamentals for Elementary School Teachers was also held on January 12.  Planning is underway for a spring meeting on the new Massachusetts Digital Literacy and Computer Science standards and opportunities for licensure featuring representatives from the Department of Education.  Planning meetings are on the third Thursdays of most months at the MGHPCC in Holyoke and online.  All are welcome.

Find out more at https://sites.google.com/site/cstawmass/

Images courtesy: CSTA WMass

Holyoke Codes Cyber Security Workshop with Girls Inc.

Holyoke Codes ran a Cyber Security Workshop for Girls Inc. on Saturday, December 8th, 2018 for Computer Science Education Week. The workshop was attended by 21 middle school aged girls, members of The Girls Inc. Eureka Program Rookies group. Eureka! Scholars spend time on the UMass Amherst campus and elsewhere, experiencing lab-based activities in science, technology, engineering, and math.

The girls learned about digital forensics and cybersecurity with topics such as cryptography, steganography, passwords, and privacy to solve a mystery in a Capture the Flag (CTF) contest.  The CTF contest was custom created by Prof. Beryl Hoffman and her Elms College Computer Information Technology students for Holyoke Codes. The contest uses a series of challenges involving cybersecurity concepts to solve the mystery of a kidnapped cat (for a demo, login with team name: test, password: test). Three new challenges were added this year, including one on Dr. Azer Bestavros’ research project on Additive Secret Sharing.

 

Support for the event was provided by the National Science Foundation CISE SaTC Frontier Award #1414119 to Boston University, entitled MACS: A Modular Approach to Cloud Security.

This experience will also be available via Holyoke Codes workshops open to the public at no charge (www.holyokecodes.org).

Pages