From petascale to exascale with Satoshi Matsuoka and Thomas Sterling

The second keynote yesterday was delivered by Satoshi Matsuoka of theTokyo Institute of Technology. He introduced us to petascale computing on the Tsubame 2.0, ranked 5^th in the world in the prestigious Top 500 and the world’s smallest petaflop supercomputer. In supercomputing, small is beautiful. “If people talk about how big their supercomputer is, they don’t know what they’re doing,” said Matsuoka. Tsubame is also one of the world’s greenest supercomputers. Compared to a laptop, it is 66,000 times faster and three times more power efficient, dealing with 44,000 times the data.

As a side note, Matsuoka told us about the discussions they had at the Tokyo Institute of Technology after the earthquake last year. The country was hit by serious power cuts in the aftermath, and the institute suggested turning off the supercomputer in the interests of saving power. The computing team pointed out that as Tsubame was much more efficient, it would be better to turn off the laptops. But of course you can’t access the supercomputer without the laptops - so in the end, they kept them both on.

Matsuoka walked us through some impressive examples of the modelling work carried out on Tsubame, such as weather prediction down to resolutions of a few hundred metres, dendritic crystal growth in lightweight metal alloys, blood flow past the type of coronary obstructions suffered recently by the Japanese emperor. Modelling earthquakes and tsunamis is naturally of high importance in the region but, perhaps surprisingly, tsunamis are relatively easy to model – it’s their effects once they move into built up areas that are more difficult to predict.

Looking to the future, Japan is setting up a High Performance Computing Infrastructure similar to PRACE in Europe and XSEDE in the US. HPCI will connect the national supercomputing centres, with the famous “K” computer as the focal point, and will be in operation by the summer 2012. HPCI will be able to boast up to 100PB dedicated shared storage.

Back in 1995, an iPad2 would have been classed as a supercomputer. Providing the next generation of exascale rather than petascale computing is going to be a knotty problem. Not least in solving how to manage the power consumption that machines with up to a billion cores will eat up.

Today’s final keynote speaker, Thomas Sterling of Indiana University talked about an execution model for HPC, moving towards the exascale. He described the “K” computer as the T-rex of the computing world, in both positive and negative senses – and not just because it has a big ‘byte’. (Pardon the pun!). He described both K and Tsubame as very impressive machines when you get under the hood, in terms of system, network and chip design. However, while there are lessons to be learnt from both of them, they are not where we will end up with future supercomputing.

Sterling’s question was where do clouds fit into the HPC future? At the moment, we would need to see a paradigm shift to broaden clouds to HPC. Machines such as the LIGO Laser Interferometer Gravitational Observatory are the source of an enormous amount of data. The largest gravitational experiment going (other than the universe itself), LIGO tracks changes in length of less than the size of an atomic nucleus in legs that are kilometres long. The challenge is to extract the signal from the noise – and there is not a lot of parallelism in the computations. Unpredictable computations of this type are not supported well by static tools such as MPI.

HPC divides naturally into three modalities – capacity computing is one, which is high throughput with job stream parallelism and without synchronisation or data exchange. Capability computing uses coupled parallel elements, has strong scaling, and is all about reducing the time it takes to do a particular, well defined job. HPC fits on the cloud via the throughput mode of computing, as it lends itself to sequential jobs of the same program on separate blocks of data, often carried out where the data is located. Sterling defined a third modality as “cooperative computing” – increasing both the speed achieved to solve a problem, but also the size of the data sets. Interim results are exchanged cooperatively to drive a dynamic computation process. It is no longer about just leveraging Moore’s Law to improve performance.

Some HPC applications are not a good fit for the cloud, such as capability codes, jobs with very large data sets, jobs with data or results that are sensitive, proprietary or security related. There are several challenges for improving efficiency and scalability, one being starvation ie not having enough concurrency, high latency which impacts efficiency, overhead affecting the critical time to manage tasks and resources and waiting, for example where there is contention for shared resources. Put Starvation, Latency, Overhead and Waiting together, and you get SLOW computing...

To bring a new clouds execution model online you need to decouple software from the hardware it runs on, be dynamic not static, hide latency and use message driven computation. “It’s not just how many cores you can squeak together in the cloud,” said Sterling,” latency matters. We need new models and programming interfaces to make this work.”

Jesse Lawrence intereviewed at ISGC 2012

We were able to drag Jesse Lawrence from the Quake Catcher Network out in to the sun and have him chat to us:

Scanning the sky with Alex Szalay

After a stunning meal at the Mitsui Japanese restaurant last night (described as ‘exquisite’ by a visiting Japanese guest, so it must have been good!), I would say that the keynote speakers this morning probably had their work cut out to grab our attention. However, they managed it effortlessly – both presentations were instantly engaging.

Alex Szalay of Johns Hopkins University talked to us about extreme data intensive scientific computing. He told us that while the big data sets attract all the press, such as the LHC and genomic research, we shouldn’t forget about the ‘long tail’ of data, generated by day to day science. Together this adds up to such a significant quantity of data, that we are actually living through a scientific revolution. Science is moving from being hypothesis driven, to data driven discoveries, and funding agencies in particular need to catch up to this trend.

Alex’s research field is astronomy, and he discussed the Sloan Digital Sky survey, which generated 100 TB of data presented through the Skyserver website, a site that has received an impressive 993 million hits over the last 10 years. Processing this data doesn’t just need computing power, but with only 15,000 astronomers worldwide, additional pairs of eyes are needed to identify the astronomical objects in the images. This ‘internet scientist’ idea led to GalaxyZoo, a mash up of the million brightest images in Sloan. To date there have been 40 million visual galaxy classifications by the public, including a few discoveries such as Hanny’s Voorwerp and the “Green Peas”.

HPC is an instrument in its own right and presents challenges many challenges: how to move data, how to look at it, how to interface to it and how to analyse it. Scientists are using HPC to solve problems such as modelling turbulence, biological systems and the brain. The Milky Way Laboratory has just been funded by the NSF, and will use cosmology simulations as an immersive lab for general users. Data Intensive Scalable Computing challenges of these type lead to a whole host of trade offs in computing. System designers need to balance the need for data storage, cost, sequential IO, fast streams, and low power – and this is a moving target that changes every year.

Storing 100 TB is really hard, especially if you need months to process the data and moving it around isn’t much easier – the quickest option can sometimes be to throw disks into the back of the car, or send it via courier if you don’t fancy a long drive. However, once you get your data where it’s supposed to be, however it travels, particularly in the cloud, you start to open up interesting possibilities that bringing small, seemlngly unrelated data into a single cloud may see new value start to emerge. This has happened in Facebook, which now has billions of photos from hundreds of millions of users. What might this bring to the long tail of science?

Making drug discovery and weather simulations more user-friendly

Yesterday afternoon at the ISGC 2012 event in Taipei, Hsi-Kai Wang of the Academia Sinica Grid Computing Centre (ASGC) presented the new web-based graphical user interface for the Grid Virtual Screening Service (GVSS).

Each year, new viruses, such as the avian flu, pose a considerable threat to human health worldwide. Drug research, however, requires a considerable amount of information and resources in addition to the time needed for development and testing. GVSS is designed to help researchers save testing time and resources by allowing them to focus on the drugs found to be effective in initial simulations.

The update also features a number of new innovative analysis and visualization tools, including a 2D ligand query feature and docking simulations based on scalable 3D molecular models. Results can be displayed in various ways, such as ranking tables, histograms, etc.

The new GUI can be accessed at http://gvss2.twgrid.org.

Another big challenge we are now facing is climate change. Running weather simulations, which can also be used for hazard mitigation, can be a very time-consuming and costly task. During a session dedicated to environmental changes and disaster mitigation at ISGC 2012, ASGC’s Yun-Pin Sun introduced another web-based GUI for gLite-based WRF (gWRF), replacing the old command-line system.

If you’d like to check it out for yourself, please go to https://gwrf.twgrid.org.

Francois Grey interviewed at ISGC

During a brief gap in his schedule Francois Grey was able to answer a few questions for us on camera about ISGC and Asia@Home.

New briefing...and hello from Stefan

February 2012 e-ScienceBriefing can be downloaded here.

February sees the launch of a special e-ScienceBriefing on networks and e-science connectivity. The vast quantities of data heralded by the looming data deluge are being dealt with through novel technologies and coordination of networks both within and across continents. But it’s a cat-and-mouse game: as new architectures and systems are built to withstand this flood of information, the need for scientists to tackle the big problems of the 21st century entails ever-larger and more data-intensive research. It’s a time of big challenges, and even bigger opportunities. Fundamental to helping us deal with some of these challenges are initiatives like the pan-European network GÉANT, daily shifting data by the petabyte to tackle everything from drug discovery to radio-astronomy, and the dawn of the Universe.

Despite all the big talk, some of the most profound benefits of e-science continue to take place on a more human scale. DECIDE, a neurological diagnostic tool that employs the high-speed research links of GÉANT to allow doctors to quickly compare brain scans is bringing better treatment to patients at a time when healthcare budgets are being squeezed. AfricaConnect, meanwhile, takes advantage of GÉANT to deliver international information networks to researchers in Sub-Saharan Africa, unlocking intellectual potential through videoconferencing and distance learning.

Science is a very human endeavour, at its most sublime when it interfaces with the social sciences and humanities, giving us information that truly enriches our understanding of our place as people. Projects like ASTRA (Ancient instrument Sound/Timbre Reconstruction Application) are making using of GÉANT to reconstruct sounds of instruments long gone through physical modeling of an instrument’s sound qualities – a boon for archaeological scientists and music-lovers alike.
It’s certainly an exciting time to be joining the eScience Talk team and wider grid community. I’d like to thank Zara Qadir for doing such a wonderful job on the latest briefing, which is being launched today at ISGC. By the way, you can read it here. If you'd like to subscribe, please click here. If you'd like to tell us how you think e-ScienceBriefings can be even better, please click here.

As for me, well…I’ve always had a passion for computers since being given a second-hand VIC-20 for my eighth birthday…and I’ve long had a passion for communicating science! My last real (research) taste of the grid was running some quantum chemistry packages on the nascent White Rose Grid (Leeds, Sheffield, York) back in 2003, and while doing a (nanofabrication/physical chemistry) PhD and working as a science communicator for the MRC, I’ve been trying to break back into the grid community. Now, that opportunity has presented itself. I’m glad to be on board!

Investigating molecular machines with Alexandre Bonvin at ISGC2012

This morning at the ISGC2012 event in Taipei, Alexandre Bonvin of the University of Utrecht introduced us to how he is using the grid to investigate the molecular machines and the social life of proteins. In many diseases, for example, it is the breakdown in communication between proteins that leads to problems. “I’m coming at this as from the user perspective,” he told us. “A key part of my role is to hide the complexity of the grid from the users.”

Structural biology uses techniques such as NMR and synchrotron X-ray analysis to delve into the structure of complex molecules, such as proteins and peptides. The big machines acquire the data from samples, and this is processed to determine what the molecule contains and the constraints on its shape. The number crunching on its structure starts there – and this is where the grid comes in. The grid is just right for this process, which is CPU intensive, consists of lots of parallel jobs but needs limited data transfer and storage.

The WeNMR project leads the largest EGI Virtual Organisation in the life sciences. It has nearly 400 users across 41 countries, including 4 outside Europe - in Taiwan in collaboration with ASGC, Latin America through GISELA and in SAgrid’s South Africa. This represents around 30,000 cores and 1.6 million jobs in the last year. Just over half of the jobs run on the Dutch National Grid Infrastructure, BigGrid, which provides 75% of the CPU hours.

As Alexandre said, grid is a complex business, and WeNMR provides a web portal offering ‘protocolised’ access to the grid for protocols HADDOCK, CS-ROSETTA and GROMACS. This is both a user friendly gateway for the NRM and SAXS communities and a virtual research platform. Linked to the ESFRI-INSTRUCT project, WeNMR works to foster the adoption and use of e-infrastructure and to get over the entry barrier reported by many users, while at the same time offering the top of the line software and services that attract expert users. The secret to sustainability is of course to get scientists, and not just the computer experts but the average user as well.

There was also a brief word on scientific cloud computing. Alexandre’s team is using cloud computing, supported by the supercomputing centre SARA in Amsterdam, to process libraries of proteins. Because this needs several sets of software and libraries, it is much simpler to set up a virtual machine once with all the software installed, then you don’t have to deal with clashes with scientific linux on the grid. “For the other portals,” said Bonvin, “grid is the perfect solution- cloud will not gain us anything.”

You can hear more from Alexandre, his team at Utrecht University and the CONCO project on the newly released YouTube video “Stories from the Grid” by EGI (European Grid Infrastructure). The first episode shows how a component of the cone snail's toxic venom is being modified using the aid of computer models, to help produce new anaesthetics and alleviate the muscle spasms caused by the condition dystonia. Episodes 2 and 3 coming soon!

Video: Alexandre Bonvin from WeNMR

Before his keynote this morning at ISGC 2012 the lovely Alexandre Bonvin from WeNMR jumped in front of or camera and give us the low down on WeNMR and why he is at ISGC this year.

The "Stories from the grid" video he refers to can be found at http://go.egi.eu/conco

Video: Daniel Lombrana Gonzalez from Citizen Cyberscience Centre

We grabbed a few minutes with Daniel Lombrana Gonzalez from Citizen Cyberscience Centre at the International Symposium on Grids and Clouds 2012

Impressive visual journey into the oceans at ISGC 2012

Prof. John Delaney from the University of Washington "WOWed" the audience in this morning's ISGC2012 keynote session with his presentation about ocean science. Holding a grid and cloud computing audience captivated for 45 minutes while not talking specifically about computers is normally quite challenging but Prof. Delaney's engaging style and amazing slides ensured nobody caught up on sleep during his session.

The Professor kindly highlighted the significant role that computation plays in enabling his remarkable research projects and reminded all those in attendance of the importance of grids & clouds in current and future scientific endeavors, specifically in oceanography.

Reminiscent of the first time I saw a Hans Rosling presentation or Al Gore's climate change lecture, Prof. Delaney used a combination of videos, photos, time lapse sequences, animations and colourful diagrams to explain complex scientific ideas and projects. If you ever get the chance to see Prof Delaney talk at future events I would highly recommend it and I would suggest that we could all learn a thing or two from his presentation style and his slide preparation. Cheers Lucien.

Video: Tom Fifield from NeCTAR at ISGC 2012

Tom Fifield is from the NeCTAR (National eResearch Collaboration Tools and Resources) project in Australia. We got a quick interview with him during a break in the International Symposium and Grids and Clouds 2012 conference

ISGC: Inclusive, Progressive, Relevant

This morning, ISGC opened with a stunning visual presentation, covering the history of the conference and the institute that runs it (ASGC). A large number of familiar faces gave their congratulations for the 10th anniversary of this excellent conference, displaying a veritable United Nations of languages and cultures. From the beginnings in networking, to the massive ramp up of data fueled by the Large Hadron Collider project, the film documented the passion of Dr Simon Lin who led the push for grid computing in Taiwan, and indeed the Asia Pacific region.

Simon Lin and Steven Newhouse then welcomed all, with a few remarks and reflections on the hard work of those attending and organising the conference, before handing over to session chair Klaus-Peter Mickel.

The first speaker, a long-time contributer who should be familiar to all, Bob Jones graced us with a poignant reminder of the things which make this conference great. He analysed its growth, in tandem with the growth of the infrastructure and concluded that ISGC continues to stand by its original aims of being inclusive, progressive and relevant.

Steven Newhouse (EGI Director) followed, to discuss the current activities in Europe. Drawing on the history theme - a short evolution of the European grid organisations led to an overview of Federated Cloud Task Force, a project designed to reiterate the success of the grid, in the cloud paradigm. EGI's cloud strategy was clear - as cloud is used more and more, many other communities not necessarily involved in today's eResearch infrstructures will begin to be. I'm sure we can all agree with Steven: "...aim to give researchers what they want, when they want...". If, in our various projects, we can make research easier to perform, have faster turnarounds and reach greater heights then we have achieved our goals.

Case Studies of Desktop Grid Particularly from the Asia-Pacific Region: Malaysia & Indonesia

This afternoon I gave my presentation regarding Desktop Grid activities in Asia-Pacific region with focus on Malaysia and Indonesia. The Desktop Grid implementation in this region may already exist but uncoordinated. The effort to introduce and coordinate the Desktop Grid was started since last year just after ISGC 2011 end. There was Dr. Robert Lovas from Desktop Grid for International Scientific Collaboration (DEGISCO) project who introduce the technology and the initiative to unite the Desktop Grid/Volunteer computing under International Desktop Grid Federation (IDGF) umbrella. Two training and seminar events was held in Universiti Putra Malaysia (UPM) and Institut Teknologi Bandung (ITB). The outcome from these events are promising: UPM has just completed their phase one Desktop Grid implementation based on SZDG-BOINC (called DG@Putra), Universiti Teknologi Petronas (UTP) and ITB about to complete their Desktop Grid implementation that UTP particularly will use SZDG-BOINC as their campus grid platform. The next step would be integrating the Desktop Grid to existing National/Regional Distributed Computing Infrastructure (Malaysia especially) and will continuously provide support (e.g. training, awareness seminar, consultation) to Desktop Grid communities in South-East Asia region particularly.

Quake-Catcher Network (QCN)

In parallel with iRODS workshop, there was another Asia@Home session. The session talked about the Quake-Catcher Network (QCN). The QCN provides software so that individuals can join together to improve earthquake monitoring, earthquake awareness, and the science of earthquakes. The QCN links existing networked laptops and desktops in hopes to form the worlds largest strong-motion seismic network. To detect seismic activities, the QCN use vibrational sensosrs called MEMS accelerometers. The QCN use volunteer distributed computing network platform (i.e. BOINC) to do the real-time computation with data from the sensor. The QCN software was developed by Stanford University and to start volunteer simply plug the seismic sensor install on their computer via USB interface, install BOINC client and attach the QCN project. The session's participants  was given opportunity to try the  sensor and the application at first hand as shown in the figure above :)

Managing data with iRODS: ISGC2012 workshop

Yesterday morning I headed over to ASGC for the Sunday workshops, which lead up to the main ISGC 2012 programme. Trying to ignore the steady drizzle and the jetlag (both my computer and body clock were telling me it was 2 o’clock in the morning), I joined the iRODS workshop, delivered by Reagan Moore of the University of North Carolina and DICE (Data Intensive Cyber Environments). iRODs, Integrated Rule-Oriented Data System, is a policy based data management system, based on open source software. It helps people to manage large collections of data, which might be scattered across multiple sites.

What I found interesting about iRODs is that it scales from managing personal data collections, such as a few thousand digital photos, up to huge international projects such as the Square Kilometre Array. If they need to, users can move hundreds of thousands of files around at a time using an automated process. iRODs is driven by a rule engine, which can configure each system differently, allowing you to choose which policies and rules to include without editing the core code.

Some examples of the types of admin policies that can be applied are file migration, format migration, distribution and usage reports. In the area of validation, policies cover integrity, replication, required metadata, derived data products, audit trail compliance, distribution and retention. Processing of data can be done either at source, or by calling routines remotely.

To try it out, Moore led the group in installing iRODs and accessing a data set in California (located not just on another continent, but still in the day before). Having done this, Moore took a quick poll of the possible use cases in the room. Possibly not surprisingly given the audience, the highest interest was in running regional, national or international data sets. However there was also significant interest in using iRODs for personal digital libraries, archives and federated collections.

The iRODS User Meeting takes place later this week on 1-2 March, and I would quite like to be a fly on the wall, just to see the range of applications in action. With data management at the front of many researcher's minds in Europe at the moment, I’m expecting to hear lots more about iRODs in the future.

Arrival in Taiwan for ISGC 2012

It was with much anticipation of the upcoming week at the ISGC 2012 that I landed this morning in Taipei, the national capital of Taiwan. Despite the cold, misty and wet conditions, the drive into the downtown districts revealed glimpses of the city's vibrancy and ongoing growth, a great setting for a conference which aims to bring together researchers and collaboration partners from across the Asia Pacific region.

Having recently joined the grid community it is events like ISGC and the associated concentration of experts willing to share their knowledge and experience that is so vital to learning the skills needed to effectively contribute to its maintenance and development.

The co-located workshops at the ISGC that I will be attending this week include the iRODS, Cloud Interoperability and Federated Identity Management workshops. Hopefully the social events and session breaks will give me plenty of opportunities to put faces to people I've only communicated with by email and to meet many new experts.

Welcome to the International Symposium on Grids and Clouds 2012

This week the GridCast team is in Taipei, reporting from the International Symposium on Grids and Clouds 2012. Already it feels like much less than a year since the last time we were in Taipei, and we are busy catching up with our friends and colleagues here at Academia Sinica and from across the region.

2012 is the 10 year anniversary of ISGC, and over the last decade it has followed how individual, innovative researchers across the Asia Pacific region have converged into a coherent community, sharing their knowledge and experiences along the way.

GridCast will be blogging from the event, alongside other delegates to bring you behind-the-scenes reports. We are also very pleased that our new Asia Pacific iSGTW Editor, Vivian Tiän-Tiän Chang will be reporting from ISGC. Vivian is based at the Academia Sinica Grid Computing Centre, and you can read about how she will be reporting from the region in iSGTW. Welcome to the team Vivian!