The interns are coming, the interns are coming…

Posted on May 16, 2013 by Vivian Benton

There’s a flurry of activity right now on the normally quiet 4th floor where my office is.  Curious to find out what’s going on, I asked Pallavi Ishwad, in charge of outreach activities for PSC, whose office is a couple doors down from mine.  “We’re preparing for the MARC interns and workshop,” she said. Oh, now I remember!  Every year in June PSC brings in students from minority serving institutions (MSIs) all over the country to participate in this program.  PSC’s MARC program, led by Hugh Nicholas, is funded by the National Institutes of Health Minority Access to Research Careers Program. The overarching goal of the PSC program is to help establish and integrate bioinformatics in graduate and undergraduate courses at these institutions.

During the summer, June 3 through August 2, eight students (four males and four females) will participate in an intense nine-week program not only to learn how to execute their research projects, or “do research,” but just as importantly, to improve non-science skills to aid in writing peer-reviewed research papers and journal articles and to be able to present scientific work at national and international conferences.

I went to the right person to ask about all this because, as it turns out, Pallavi is taking a more active role in helping to mentor the interns. The program’s mentors make sure, among other things, the interns stay on track with their projects.  Once a week the students have to present their progress to the staff. It’s during this time problems can be caught early on and students can refine, if necessary, their goals for their specific projects.

I was curious to find out what projects these interns would be working on.  I was even more curious about how their projects were selected. Well, the specifics vary and I admit my eyes did glaze over when I asked the project’s co-investigator Alex Ropelewski, and he mentioned RNA, Trinity and guinea fowl in the same sentence.  TMI for this blog! I did, however,  glean from our conversation that Trinity is the name of a software program some of the interns will be using to assemble their sequence data.

Throughout the year, leading up to this summer program, Alex, Hugh, Pallavi and other project staff travel to and work with liaisons at five MSIs—Tennessee State University, North Carolina A&T State University, Johnson C. Smith University, University of Puerto Rico at Mayaguez, and Jackson State University—developing bioinformatics under the grant. The summer internship program is open to students at other MSIs as well and this year includes interns from the University of Texas at El Paso and Morgan State University.

One example is a project being researched by incoming intern Aryan Vahedi-Faridi from Morgan State University. He has data from an experiment that compares the effect of sleep deprivation stress on gene expression in the brain and hopes to analyze this data during his internship.  He wants to understand how the different signaling pathways are affected by this type of stress and relate these findings about sleep deprivation and perhaps treatment of these symptoms. Another intern, Kayla Hinson from the University of Texas at El Paso, wishes to expand her work on population biology and genetics of freshwater invertebrates. Her focus will be on arsenic detoxification pathways and identifying potential detoxifying enzymes in an aquatic model organism. You get the idea. This is serious. And intense! And I now know way more than I intended about correlation values, spicing variants and microarrays.

I’ll be following along this summer to see how they are progressing. I might even attend one or two of the weekly sessions so I can report back to you later on how the program is going. So stay tuned!


Posted in Bioinformatics, General, Outreach, People, Training | Tagged , | Leave a comment

Science Wednesday: Sometimes, You Need Gin to Spell Synergy

Posted on May 1, 2013 by Ken Chiacchia

Hepburn and Bogart delivered more than the sum of their talents in "The African Queen" -- so must the stakeholders in VECNet, to conquer malaria.

I’m just gonna fess up and admit that the story I’ve been telling the PSC VECNet Cyber-Infrastructure crew is probably a myth. Director John Huston and Humphrey Bogart probably didn’t stave off malaria while filming “The African Queen” on location in Uganda and the Congo by keeping themselves blotted on gin and tonic.

It has some verisimilitude. Bogart and Huston, famous drinkers, did avoid contracting the disease. And tonic water contains quinine, the world’s first effective pharmaceutical malaria preventive.*

But the story falls apart on closer examination. Bogart, Wikipedia tells us, credited dodging the dysentery bullet to “the large supply of whiskey he had brought along with him.” Anjelica Huston, too, quoted them as drinking Scotch, not G&Ts: “Whenever a fly bit Huston or me, it dropped dead,” Bogart later told her.

Never mind all that. The more interesting part of the story, at least from the point of view of VECNet and its CI project, is that the incomparable Katherine Hepburn, disgusted by the boys’ functional alcoholism, eschewed the booze for local water — and came down with dysentery as a reward for her virtue.

Hepburn and Huston didn’t get each other at first. Vastly different people, these two geniuses needed some time to work around their different ways of looking at the world, of working. It’s our gain that they stuck it out; the combination of their talents produced something epic, vastly greater than its already considerable sum.

So, too, VECNet. What our PI, Nathan Stone, and collaborator Greg Madey, PI at the University of Notre Dame, are aiming for is nothing less than a plug-and-play simulator that allows people with minimal IT training to ask the question, “What happens to malaria incidence if I do X?” and get a statistically valid answer. One of high enough quality to make real-life decisions on what research to conduct, which products to develop, what policies to adopt, and whose projects to fund. And, importantly, to identify which other VECNet users have other pieces of the puzzle, spurring collaborations.

“The larger VECNet community has been at it for over three years now, assembling some of the necessary models and data,” Nathan tells me. “This year’s challenge is all about integration and execution — like assembling a race car with parts from two Humvees and a bus so that any cabbie could use it to win the Indy 500.”

It’s audacious. It’s promising as all hell. And it’s going to take input from people who see the world — and how to attack the problem — very differently:

  • Scientists, famous for not wanting to take a step without the other foot firmly planted, will need to see the project as built on solid scientific footing, with transparency as a chief goal.
  • Engineers, who tend to view the world in terms of what the product does rather than the process of building it, will want to see a system that above all works.
  • Corporate researchers will need to have their proprietary ideas guarded, so that they can reap the fruits of their R&D labors.
  • Government officials will need answers that are both economically feasible and politically doable in the context of their own systems of governance.

I don’t think I’m giving away state secrets to say that these folks are going to be — have been — arguing over how best to do all of the above. But it will be to the world’s gain that they stick it out.

______________________________

*Modern tonic water has far less of the stuff than the originals, and so wouldn’t work.

Posted in Bioinformatics, General, HPC Research, People | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , | Leave a comment

Nathan Stone: Physicist Turned Malaria Warrior

Posted on April 30, 2013 by rscibek

Nathan in his office.

I’ve been working with Nathan Stone for almost ten years, and over the past few months there has been a question nagging at me.  How in the world does a guy who has a Ph.D. in Experimental Nuclear Physics wind up being involved in PSC’s newest collaborative project, VECNet, which is researching new strategies to better control/eliminate malaria in developing countries?  In my mind, experimental nuclear physics != controlling malaria! So I sat down with Nathan and got a little more insight as to what landed him his current role of Modeling and Project Manager for the VECNet Cyber-Infrastructure (CI) team.

Nathan describes his undergrad major as a happy accident.  When he applied to Purdue University as an undergrad, he wanted to be an engineer. Unfortunately, when he applied it was too late to get into the engineering school since all of the slots had already been filled.  So he applied to the College of Science at Purdue and hoped to transfer to engineering his sophomore year.

As it turns out he stuck it out in physics.  He even participated in a REU program at Michigan State between his sophomore and junior years.  While working in the accelerator lab on campus he realized that research was much more than writing a book report.  He enjoyed his research experience at Michigan State and in the accelerator lab so much that he went back to get his Ph.D. there.

Nathan giving a virtual machine room tour at PSC's 25th anniversary celebration to a group of K-12 teachers.

After completing his post-doc at Berkeley, Nathan started working at Brookhaven National Lab on Long Island as a full time staff scientist.  He was a member of a collaborative team to bring up a unique new particle detector, STAR.  Nathan was involved in data calibration and acquisition and experiment controls for the new device.  Before STAR could receive funding to build the device, they needed to prove to the funding agency, the Department of Energy(DOE), that it would work.  This is where PSC enters the story… the STAR team ran simulations on PSC’s supercomputers to make their case to the DOE.

The STAR project required collaboration meetings about every 6 months, and during one of the meetings, Nathan gave a facility tour to PSCers Mike Levine & Sergiu Sanielevici.  Sergiu was looking for staffers to participate in the National Institutes of Health’s Collaboratories project.  Participants would help find new hardware, software, techniques, and tools that could help geographically distributed collaborators to work more effectively together.  Nathan’s experience with various partners at Brookhaven made him a perfect fit, and Mike & Sergiu encouraged Nathan to apply at PSC.

Nathan’s tenure at PSC has involved a number of projects, all of which have had two commonalities, ensuring that PSC users and collaborators have remote, secure and fast data access, and working on and finding tools to make collaborative projects more efficient and effective.

The Star Trek door chime mounted outside of Nathan's office that makes a swoosh noise everytime you visit.

So when PSCer Shawn Brown needed developers with a mind for research for the MIDAS influenza modeling project, he asked Nathan to join his team.  Nathan worked to solve some data issues with the FRED model that they were using and then he moved on to working on developing a new vector model for mosquitoes to better track and predict spread of the Dengue virus.

Later, Bruce Lee at Pitt, who had done some economic modeling for the VECNet project, got Shawn and Nathan involved.  The VECNet project needed a cyberinfrastructure that would help further the research, but they just hadn’t been ready to justify it.  Nathan did an applications analysis, which had not been done before, to justify and project the need for a more complete cyberinfrastructure for the project.  Proposal writing and contract negotiations took over a year, including traveling and meeting with various project collaborators and end-users (a.k.a. “stake holders”).  All of the hard work paid off recently when a generous grant of $1.6 million was awarded to make this proposed cyberinfrastructure a reality.

Now, Nathan is the project manager of the CI team and is ensuring that all of the pieces come together to help provide the VECNet team with tools that will make their research and collaboration more effective and efficient.

When Nathan’s not trying to save the world, one mosquito at a time, you can find him spending time with his family. He has 5 kids: two boys and three girls.  He also enjoys hunting, and driving his co-workers mad with his Star Trek electronic door chime.

 

 

Posted in General, HPC Research, People | Tagged , , , , , | Leave a comment

Science Wednesday Thursday: Saving Sergeant Chiacchia

Posted on April 11, 2013 by Ken Chiacchia

“The July 12, 1973 fire at the St. Louis National Personnel Records Center (NPRC) destroyed approximately 80% of Army personnel records from 1 Nov 1912 to 1 Jan 1960; and, 75% of the Air Force records from 25 Sep 1947 to 1 Jan 1964. In all, between 16 to 18 million military service files, including those for WWI and WWII, were destroyed.”

—   Kathleen Brandt, archives.com

The 1973 NPRC fire is of more than passing interest to me. Among the millions of Americans who lost records — 80 percent of those who served in the Army in both World Wars and Korea — was Alfred Bernard Chiacchia, my grandfather.

“Pete” Chiacchia — I never heard anybody but my grandmother call him “Alfred” — arrived in Normandy somewhere about 25 days after D-Day. Luckily, Pete — a 35-year-old draftee with three kids — missed the fighting there. But as a combat engineer with George S. Patton’s Third Army, he witnessed as much as a single dogface could have seen of Patton’s sweeping advances through France, Benelux, and into Germany. He told a story about the Battle of the Bulge that, to my deepest regret, I never heard — but which family members assure me was as hilarious as it was epic.

I’m not naïve. War is war. It was never really clear what portion of his stories was real and how much was … enhanced. But he always made it a good story.

“They made me a sergeant,” I did hear him say. “Twice.”

I think it’s safe to say that I inherited the storyteller gene.

Ms. Brandt assures me that, given his birth and death dates, Social Security number and, possibly, unique service number, I can probably reconstruct those records from other sources. I may well take up that quest — not that I’m certain how to get some of that information, now that his generation in my family is all but gone. But it certainly would have been easier just to get them from the NPRC.

Which brings us to our topic of today (well, yesterday, if you want to get technical): meaningful digitization of hand-written and other non-electronic records.

I say “meaningful,” because it’s pretty easy to digitize records if you don’t intend to make use of them. Moving from visual scans to searchable data, however, is an incredibly time- and manpower-expensive prospect, taking thousands of hours by human workers to recognize and transcribe data that were recorded in cursive script into digital form.

Among those working to change that are Liana Diesendruck, Luigi Marini, Rob Kooper and Kenton McHenry at the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign. They’ve been using PSC’s Blacklight to analyze scans of the 1940 Census to try to improve the ability to automate the process.

I look forward to telling this story in detail in the near future. Part of the beauty is that their approach doesn’t try to reproduce humans’ ability to recognize the script, which is beyond computers’ capabilities today. Instead, they match entries based on how similar they are in shape. So one Census worker’s “John” can be matched with another’s, even if their handwriting is very different. Humans can choose from the possible matches — and the system learns from their choices, getting better. Ultimately, the goal will be to digitize many different types of records, so that they won’t merely be protected from loss — they’ll be easily accessible, searchable, living documents available to generations to come.

It’s a story with a happy ending. Pete would have liked it.

Posted in General, HPC Research, People, XSEDE | Tagged , , , , , , , , , , , , , , , , , | Leave a comment

Ken Hackworth: The Man, The Myth, The Legend

Posted on April 3, 2013 by rscibek

Ken working hard at his desk.

We haven’t talked about any of our fabulous staffers in a while and since we did two posts on the XSEDE XRAC process a few weeks back, here and here, we thought this would be the perfect time hear about the man behind it all,  PSC staffer and the XSEDE Allocations Manager who isn’t afraid to wear pink, Ken Hackworth.

Ken as a student, sporting the striped shirt.

Ken has been working for PSC since almost the beginning.  We were founded in 1986 and Ken started working for the Center as a help desk student in 1987.  Around here our students get pulled into a variety of activities, and Ken’s experience was certainly no different, since he also helped out with allocations and as a grunt work office assistant during his tenure as a student.

When Ken graduated from the University of Pittsburgh in 1988 with a B.S. in Economics and a minor in Computer Science, he was hired full time at the Center as a User Consultant.  He assisted our user community by answering their questions and solving any problems that they had. In 1999, he became our User Relations Coordinator which had him more involved in allocations, consulting and overseeing the students working in the hotline and allocations. He also took on coordinating PSC’s booth presence at the annual Supercomputing Conference.

Ken golfing in Nevada prior to the March XRAC meeting.

Ken got his ‘big break’ in July of 2011 when he took over as the XSEDE Allocations Manager.  Ken’s experience here at PSC made him a perfect fit for the job. As the XSEDE Allocations Manager, Ken does a lot, including overseeing the XSEDE allocations process, coordinating and organizing reviewers and their recommendations, working with NSF to incorporate and develop guidelines and of course working with XSEDE service provider site representatives.

When Ken isn’t in the office you can most likely find him on the golf course or spending time with his family.  He’s been married to his high school sweetheart for 23 years, has a 12 year old daughter, and 2 dogs, a black and chocolate lab. Ken also reports that among his many superpowers, the most noteworthy is a photographic memory of golf courses.

Posted in People, XSEDE | Tagged , , , , , , | Leave a comment

Science Wednesday: Instrumentation and Insight

Posted on April 3, 2013 by Ken Chiacchia

I’m not a big fan of second-guessing the ancients’ perception of natural events. People in earlier times didn’t have incorrect ideas of what the world was like because they were ignorant — let alone stupid. Generally, their ideas were pretty much in line with the phenomena they saw, given the quality of the instrumentation they had to measure those phenomena.*

Case in point: Archaeologists have (re)discovered the gates of Hell. The literal gates of Hell. Quite frankly, even knowing all about carbon dioxide and other toxic gases emerging from volcanic caves, it’s hard not to side with the ancients on this one. You watch birds dropping dead at the entrance of an opening into the bowels of the earth, given no volcanology (the word hadn’t been invented yet), no chemistry, no physiology. Of course you’re going to build a temple and bring sacrificial animals (carefully) to the gate, saying, “Please not me, not yet.”

Which is all by way of saying that these things look pretty clear with a couple thousand years of hindsight. And if you improve the instrumentation, you improve your perception.

I’m looking forward to talking next week with Elia Zomot and Ivet Bahar at Pitt about their paper from last month showing how a molecular door in nerve cell membranes uses the relatively high concentration of sodium ions outside the cell to drive neurotransmitters inside. Kind of like an oshiya in the Tokyo subway, cramming commuters into a car. Important, because getting rid of the transmitters in the junction between nerve cells helps turn off neural signals when they’re no longer needed. That process may be faulty in conditions like epilepsy, ischemia and other stroke-associated causes of nerve damage, and Huntington’s disease.

For this task, the researchers used Anton, a PSC resource from D. E. Shaw Research that specializes in simulating the motions of all the atoms in large biomolecules. Anton allowed them to run an electronic model of the door, a protein called an aspartate transporter. First as it opened to the outside to pick up an aspartate molecule, and then as its components pivoted, clothespin-like, to release it inside the cell.

Clothespin-0157e3

Fair-use image from Wikimedia Commons, http://commons.wikimedia.org/wiki/File:Clothespin-0157e3.jpg.

Their model spanned multiple microseconds. That’s a long stretch of time in molecular dynamics. It also would have been a major computational challenge as recently as three years ago, before a D. E. Shaw Research grant made PSC’s Anton the first of its kind available to researchers outside the company.

Based on what I glean from the paper, the “old,” static images of the transporter that we had from previous crystallographic experiments had given the impression that the door opens differently than the Anton calculations show that it has to move. A subtle correction, but one that could pay off in addressing neural damage in a number of diseases.

Our own Markus Dittrich, director of the National Resource for Biomedical SuperComputing at PSC, tells me that Anton can run these kinds of simulations about 100 times faster than anything else out there. Anton handles microsecond-scale simlations with relative ease, offering even glimpses into millisecond timeframes. By comparison, the brain can perceive auditory events as short as about 10 milliseconds.

That’s amazing, even if it isn’t quite as profound a transformation as explaining cave-entrance deaths as a phenomenon of carbon dioxide poisoning rather than of the miasma arising from the land of Death. But then, they had over 600 times as long to work on the problem.

*******************************

* I’m not counting the pre-Galilean notion that heavier objects fall faster than light objects. Then as now, people can hold wrong ideas about things they haven’t bothered to measure.

Posted in Bioinformatics, General, HPC Research, People | Tagged , , , , , , , , , , , , , , , | Leave a comment

Science Wednesday: The Inescapable Logic of Graphs

Posted on March 27, 2013 by Ken Chiacchia

Yahoo News recently had an amazing graph analytics page (see our page on Sherlock, and this iSGTW article, for more on PSC’s recent activities in the field) concentrating on the voting records of U.S. senators. Specifically, how often they voted together. Here’s the embedded graph:

Note that the following is an honest attempt to look at the issue in a non-partisan way. (Impartiality, to my mind, is a myth; instead we should strive to be fair.)

The graph opens up at 65 percent — that is, every pair of senators who voted together 65 percent of the time in the 113th Congress is linked by a line (in the parlance, “vertices” linked by an “edge”). The gist of the article accompanying the graph is that the Democrats’ party unity is slipping, with more Dems breaking from the pack rather than voting in unison. That may well be, comparing the 113th to earlier Congresses. But it isn’t really what jumps out at me.

At 65 percent (note the slider at the top, which allows you to change the percentage), we see pretty solid party unity — the Dems are in a little ball, representing their voting solidarity, as are the Republicans. But we do see some interesting deviations.

Frank R. Lautenberg, a Dem from my home state of New Jersey, is a bit of a mystery. (You can identify the individuals by left-clicking on the little dots that represent them; annoyingly, if you want the greying and pull-down menu to go away once you’ve done that, there’s no “deselect” feature — you have to re-load the page.) He’s voted zero times with anybody at 65 percent. He doesn’t connect with anybody else until you slide the percentage down to 25 percent; apparently, he’s voting exceptionally independently. I haven’t really dug into his positions, but it’s interesting that the so-called independents — Angus King, I-ME, and Bernie Sanders, I-VT, track lock-step with the Dems.

Now Bernie is a socialist, so maybe it’s no surprise that he votes more with the not-quite-left-enough rather than with the right. King is a little more interesting, since Wikipedia says he justified caucusing with the Dems to his constituents as a practical matter of having more clout by being with the majority party — and that he’d caucus with the Republicans if that changed. His voting record seems to call that into question, even if he was absolutely sincere at the time.

It’s more subtle, but at 65 percent you can also detect two red dots that pull slightly away from the Republican pack, toward the Dems (though far closer to the Republicans). That’s Susan Collins, R-ME, and Lisa Murkowski, R-AK, two moderate Republicans sometimes called RINOs. Once again, the voting record tends to question that assignation, at least at this level of voting agreement.

Move the slider to 75 percent, and things start to get interesting. Now Joe Manchin, D-WV, starts to pull out by himself. Though he still has a tenuous connection with his party, his conservative-independent streak is starting to show.

At 80 percent, we see an interesting phenomenon — Collins and Murkowski are now off in their own little clique (the technical term for this phenomenon), voting with each other 80 percent of the time but nobody else that often.

This reminds me of a graph that Nick Nystrom showed me just before his interview with Bill Flanagan on Sherlock. The graph was based on the insulin response (for my sins, this was the topic of my dissertation). Cliques appeared on this one as well — little clusters of genes linked with each other but only connected with the main ball of interactions by a narrow isthmus. I sensed that this was biologically interesting, but it wasn’t until later that it hit me: These represented biological activities that the insulin response shares with that of other messengers — think glucagon, maybe, or epinephrine. Similarly, I think we can hypothesize that Collins and Murkowski, at the level of 80 percent, are sharing a political philosophy that, while connected with the GOP at a looser level of association, nevertheless is distinct and shared.

At 90 percent, we see the reason for my sort-of-dismissal at Yahoo’s contention that the Dems are fraying in a major way. At this level, Dems are starting to peel off, with some disconnected from their party and anybody else. This is the phenomenon the Yahoo analysts were noting; but you also have to look at the Republicans, who now form a crescent, spreading out in a kind of penumbra of the still-cohesive Democratic ball. It’s clearly still a party — but (again, at the 90-percent level of agreement) it’s a party only by the property of transitivity, and clearly has lost its cohesion.

These aren’t judgments — they’re observations, driven by graph analytics.

Now I will edge out onto the diving board, and make a prediction: The Republicans’ time out in the cold, much like the Democrats’ in the ’80s, won’t end until that crescent starts to coalesce into a ball. Or — though this wouldn’t be fun for either party — until the Dems spread out into their own crescent. I leave it to the reader to decide which would be best for the American public.

Posted in General, People | Tagged , , , , , , , , , | Leave a comment

XRAC Meetings aka Quarterly Gorge Fests (Part 2)

Posted on March 22, 2013 by rscibek

I’m back with the follow up from last week’s post XRAC Meetings a.k.a Quarterly Gorge Fests (part 1).

Some of the delicious breakfast food we enjoyed.

Monday morning starts early… with breakfast available beginning at 7 a.m. The breakfast buffet usually includes pastries, eggs, bacon, sausage, yogurt, fruit, and sometimes grits and French toast!  Reviewers and XSEDE site representatives trickle in to wake up with their coffee, eat and get settled until 8 am which is when the meeting starts.

The entire group stays together for the morning to hear presentations of the proposals falling into the Plenary review session.  What constitutes a Plenary proposal you ask? Plenary requests are the big requests for the current request period and they typically include 10 million or more service units.  However, service unit cut off can fluctuate from meeting to meeting, but 10 million is where it sits right now.

Review board and site representatives during Plenary session.

This is also when the meeting really starts to get interesting, since the reviewers are adequately caffeinated at this point and proposals are presented to the entire board.   The reviewers take their review responsibilities very seriously and really try to be fair and honest with their reviews. The length of each presentation varies, but I can tell you as an observer, that the good proposals require less discussion, because everything that the reviewers need to make their recommendation is included. What constitutes a good proposal you ask? A good proposal is well written, includes a clear computation plan, clear justification of the requested resources, detailed scaling data, and uses efficient use of the requested resources.

Morning break snacks

Once we get through the Plenary proposals, we usually take a short break, and split into two smaller groups that are more science specific for the “smaller” requests.  This break of course also includes more food – usually snacks appropriate for 10 a.m. At this particular meeting we had fruit, assorted yogurts, granola, trail mix and fruit Napoleons & assorted mini fruit tarts.

I’d like to say that the smaller groups move faster through the proposals but that isn’t always the case, since they evaluate all of the proposals the same way.  We continue presenting and evaluating proposals until lunch time, when we take a much needed break.

Fettiucini alfredo & baked rigatoni

At this meeting we had a little Italy themed lunch complete with antipasto platter, Caprese and Ceasar salads, chicken Vesuvio, baked rigatoni, beef Pizzaiola, fettuccini alfredo, vegetarian lasagna and of course assorted mini Italian desserts.

Once the lunch break is over we continue on in our smaller groups.  Around 2pm there’s an afternoon snack. We’ve eaten 4 times during the course of 7 hours! Now do you understand why this meeting is a ‘Gorge Fest’? We enjoyed soft pretzels and Ben and Jerry’s Ice cream bars.  Depending on how many proposals there are and how long the discussion goes in each of the groups we can usually wrap things up around 3 pm.  This is where the reviewers jobs end and the allocations team really starts.

For each request period each XSEDE site submits a total of how many service units are available.  The allocations team, led by Ken Hackworth, have to take a look at the total awards for each XSEDE resource and determine if any additional adjustments need to be made.  When the reviewers make their award recommendations, they do it purely based on the proposal, not on available services units. So at the end of the meeting there could be a resource that has a total recommended award totally over the amount that is available.    This actually happens more often than it doesn’t. So the allocations team spends some time making sure that none of the resources are over allocated and the researchers get as much of the reviewer recommended service units as possible.

All you can eat...

This process usually takes us into the late afternoon early evening.  At this point most reviewers have left to catch their flights home, and all that’s left is the allocations team and a few reviewers that didn’t have to go home right away. So what do we do once we’re done with the meeting you ask?  Well, we eat again.  We usually venture out of the hotel and for this meeting we took it to an whole new level. Since we didn’t quite have enough to eat already we went to Texas de Brazil.  Like we really needed to top the day off with an all you can eat salad bar and o, so many meats and preparations.

Well that pretty much sums up the Gorge Fest, and more importantly the XRAC Meetings.

Posted in General, HPC Research, People, XSEDE | Tagged , , , , | Leave a comment

XRAC Meetings aka Quarterly Gorge Fests (Part 1)

Posted on March 15, 2013 by rscibek

 

Las Vegas, NV : New York Hotel & Casino, Las Vegas, NV

New York New York Las Vegas Hotel and Casino photo courtesy of city-data.com

We just got back from the March XSEDE Resource Allocations Committee Meeting (XRAC) that was held at New York New York Las Vegas Hotel and Casino.   These quarterly XRAC meetings are where allocations requests are approved or in some cases denied, for each allocation cycle. I’m sure many of our XSEDE users wonder how and why the allocations decisions are made regarding their specific requests.  Well have no fear, I’m here to tell you exactly what goes down in detail.

Our meetings typically start on a Sunday night with a small reception consisting of about 30-40 reviewers, and about 10-15 XSEDE site representatives and allocations staff.  Reviewers usually sign on for three year review terms, so this reception is a great time for us all to meet the new guys or girls that have joined the review board that quarter.  More importantly, this is also where the Gorge Fest starts.  Usually while we’re milling around chatting, and enjoying a tasty beverage, there are appetizers to hold everyone over until dinner.

A portion of Sunday night's buffet selection.

Immediately following the reception, we’re ushered into a meeting room that features a full buffet.  The buffet differs for each meeting, but since this quarter’s meeting was in a New York themed hotel, the buffet was a New York City tour with Hot Dogs, falafel, General Tso’s Chicken, Penne Alla vodka, Catch of the Day, Beef Brisket, and an array of salads and desserts.

We could show you who the reviewers are, but then we'd have to reject your proposal!

Once we’re all adequately stuffed we move into a different room where the meeting really starts.  Usually, it begins with a series of presentations from the various site and NSF representatives outlining any new resource updates or policy changes for this allocation period.  Once everyone is up to speed, we move onto one of the most important parts of the entire process, the caucus. Prior to the meeting reviewers are assigned a number of proposals to read and review before they arrive.  Depending on the type of proposal, it could have anywhere from 2-4 reviewers assigned to it.   During the caucus, the reviewers for each proposal get together and discuss the merits and in some cases demerits of each proposal.  They come to consensus, and prepare their recommendation(s) to present to the entire review board the next day.  This can take a while and reviewers are often up late to make sure they can caucus for each proposal.

It’s usually after 10 or 11 p.m. once all of the reviewers have trickled out of the room and the allocations staff calls it quits for the night.  I’m starting get hungry again, with all this talk about food, so I’m going to grab a snack. Stop back next week for part 2 of the XRAC meetings, aka Quarterly Gorge Fests recap.

Posted in General, HPC Research, XSEDE | Tagged , , , , | Leave a comment

Training Wheels

Posted on March 5, 2013 by Ken Chiacchia

Tuesday, March 5, 2013, 14:01

Yesterday the University of South Florida announced that their faculty members and collaborators have designed a material that may  be able to capture carbon dioxide from a smokestack. They did this, in part, with the help of molecular modeling on PSC’s Blacklight.

Now there’s a pretty big leap between a Nature article — as significant an accomplishment as that is — and a working scrubber that you can install in a power plant, or a car exhaust. A lot of work remains to be done, and there’s a chance this material will need to be much improved, or even wind up being a dead end. But right now it’s a big breakthrough, because it works in the presence of water vapor. That’s just about a universal component of factory and car exhaust, and had been a real problem for previous candidate scrubber materials.

Carbon sequestration is often portrayed as the Great Hope for preventing global climate change. Or — due respect to our colleagues at USF and King Abdullah University of Science and Technology — the biggest boondoggle next to denialism, depending on whom you ask. Of course, it probably will prove to be yet another incomplete but necessary measure for us to get on top of what may be the biggest problem — and problem with denial — of our age.

Let’s face it; we’re overloaded on doom and gloom — even among those who accept the science, it often seems hopeless even to get started on a solution. No less prestigious a personage than James Lovelock has floated the ludicrous idea that we need to suspend democracy to prevent a climate change disaster. I can see where he’s coming from. It does seem like mitigating climate change will involve political will we can’t muster, money we don’t have, and, worst of all, changing our individual behavior. Preventing any action at all, on the other hand, only requires a script kiddie and a poorly protected email system.

I know what you’re thinking: Postponing the shift from fossil fuels to renewables is itself a delaying tactic, and poses its own drawbacks. Brother, you don’t need to lecture me:  I can see the fracking tower from my driveway, lit like — ah, fracking — Minas Morgul* and shrieking about as loudly. I know that if those guys don’t do their job right, it could be the end of our sweet little well — our only water source.

On the other hand, the water is still good after weeks of drilling. Maybe they will do it right. Maybe they’ll pull up cheap, scrubbable fuel that will be our training wheels until we can develop the technologies that make fossil fuels … fossil fuels. The stark fact is that, decades into realizing we’re in trouble, we’ve done next to nothing about it. And that’s not only because of denialism. It’s also because there are no attractive alternatives yet, and it looks a lot like actually bringing the CO2 level down is going to require changing how we live and all sorts of other unpalatable stuff.

In that context, don’t under-estimate the value of training wheels.

_________________

* I’m a little color blind — it may not be glowing green. But it is glowing.

Posted in HPC Research, People, XSEDE | Tagged , , , , , , , , , , , | Leave a comment