Category Archives: Uncategorized

Final Conference of 2014 – HPCAC South Africa Conference

Held in the famous Kruger National Park in South Africa, the last HPC Advisory Council conference for 2014. Many thanks to Brian Sparks for managing the event, as well as to Prof. Dhabaleswar K. Panda (Ohio State University), Hussein Harake (Swiss Supercomputing Centre), Dr. Robert Sinkovits (San Diego Supercomputing Center) and Addison Snell (Intersect360) for their presentations. 2015, here we go…



HPC Music

HPCmusic is an advanced research project about High Performance Computing and Music Production dedicated to enable HPC in music creation. It’s goal is to develop HPC cluster solutions that further enable the future of music production.

During the last 10 years, with the advancement of HPC, it gradually became possible to create a cluster that can facilitate the musical size of a full orchestra and let the composers change the ‘notes’ and the ‘feel’ of the music in real time. No other technological solution ever managed to achieve this.

We now have the chance to set the standards on how a cluster should work in order to serve composers on creating music inside a dedicated virtual environment. The goal is to provide the composer the ability to incorporate all the qualities of a symphonic orchestra, the improvisation of a jazz band and repeatability and consistency of a hybrid analog/digital system.

Antonis Karalis research to date clearly indicates that HPC projects are the only way to move music forward. With low latency systems that become transparent to the composer we increase the potential of new musical ideas.

You can read more by looking at the HPCmusic subgroup under the council subgroups section.

Below Antonis Karalis working on a cluster at the HPC Advisory Council cluster center.


Gilad, HPC Advisory Council Chairman.

The 2nd GraphLab Workshop – Update

The main aim of the 2nd GraphLab workshop is to bring together researchers from academia, as well as data scientists from industry with the special focus of large scale machine learning on sparse graphs. We have secured talks and demos about the hottest graph processing systems out there: GraphLab (CMU/UW), Pregel (Google), Giraph (Facebook) , Cassovary (Twitter), Naiad (Microsoft Research), GraphBuilder (Intel Labs), Grappa (UW), Combinatorial BLAS (LBNL/UCSB), Allegro Graph (Franz) ,Neo4j (Neo Technology), Titan (Aurelius), DEX (Sparsity Technologies), YarcData and others!

The workshop will take place on Monday July 1st at the Nikko Hotel SF. Workshop website:


Dr. Danny Bickson

Carnegie Mellon University

The 2nd GraphLab Workshop

The 2nd GraphLab workshop ( brings together researchers from industry and academia to discuss challenges of applied large scale machine learning, with the focus of sparse graph algorithms. It is a single track one day event to be held on Monday July 1st in San Francisco.

We are going to have talks and demonstrations of the hottest graph computation frameworks:

GraphLab (UW + CMU), Giraph (Facebook), Pregel (Google), Cassovary (Twitter),

Computational BLAS (LNL) and others.


HPC Advisory Council members can receive 15% discount – please contact for more info.


Dr. Danny Bickson

Carnegie Mellon University

HPCAC 2012 Spain Conference

Last week we had our first Spain conference, organized together with the Barcelona Supercomputing Center. The conference held in the Malaga University, Malaga. Malaga is a great place to visit in the summer if you like good food and enjoy the sea. And of course, if there is a HPC conference nearby…. :-)

70 attendees joined us for a full day of interesting presentations from great speakers. You can find all of the presentations at InsideHPC covered the day and you can find some videos on

I want to thank all of our sponsors – AMD, Dell, HP, Mellanox, Panasas and Boston. Without them we could not have the event.

Our next conference is in China – October 28th!



HPCAC Spain 2012

ISC’12 Student Cluster Competition

At ISC’12 the HPC Advisory Council managed the first ISC-HPCAC Student Cluster Competition. Five universities competed on achieving the best applications performance under a tight power budget of 3000 Watt – Karlsruhe Institute of Technology (Germany), The University of Colorado (USA), Stony Brook University (USA), Tsinghua University (China) and the National University of Defense Technology (China).

The competition was covered by InsideHPC (, Gabriel Consulting Group ( and The Register ( The competition was also covered by several Chinese media channels, including CCTV (morning news coverage on CCTV was broadcasted on 6/21

Here are the competition results:

The first set of awards was given as a recognition for specific achievements that the teams have demonstrated throughout the three days of the competition: 

– KIT (Germany): the highest Linpack result on x86 based cluster

– StonyBrook University (USA): for impressive troubleshooting and overcoming hardware and software challenges

– Tsinghua University (China): for understanding the impact of weak and strong scaling on application performance

– Colorado University  (USA): for being first to run all applications, including the secret ones

– National Defense of Defense Technology (China): for building the highest core count system in a hybrid architecture.

The three major awards:

– The award for the highest Linpack score of 2.65 Teraflops was presented to the National University of Defense Technology, China.

– The award for the “Overall Winner” which achieved the highest aggregate points total for all the benchmarks included in the competition was presented to Tsinghua University, China

– The “Fan Favorite” award for the team that got the most number of unique votes from ISC participants during the competition was presented to KIT, Germany



The HPC Advisory 2012 European Conference

On June 17th we held the HPC Advisory 2012 European Conference in Hamburg Germany, the day before ISC’12. Around 100 attendees participated in the workshop. We would to thank AMD, DataDirect Networks, Mellanox, Xyratex, Intel, Transtec, 451 Research, HPC Wire, InsideHPC and Scientific Computing World for sponsoring and covering the conference. We would like also to thank the presenter for their interesting presentations. One can find all of the presentation on the conference web page –




Colleges & Supercomputers: State, Schools and Private Sector Teaming Up To Push Innovation

The role of technology in education is greater today than ever before. As a result, many cutting-edge colleges and universities are teaming up with state agencies and private businesses to deliver state-of-the-art computing systems to students. These collaborations enable researchers and technicians to advance innovation at the college level.

One such partnership took place in May 2011, when IBM donated a supercomputer to New York’s Union College to celebrate the school’s newly opened Peter Irving Wold Center. A $22 million project, Times Union reported that the center is devoted to “innovative research and exploration”. Now, thanks to IBM, researchers at the center will have access to the $1 million ‘intelligent cluster’ of 88 servers and more than 1,000 processors, which collectively allows 10 billion calculations 1,000 times per second.

IBM reportedly plans to work with Union researchers to develop a neuroscience-imaging project and conduct an environmental analysis of the Mohawk River. John E. Kelly III, a Union alumnus and IBM’s current senior vice president and director of research, noted that the system is a fitting counterpart to the supercomputer found at nearby Rensselaer Polytechnic Institute. He said, “This concentration of computer technology and computational horsepower between RPI and Union is incomparable anywhere in the world.”

Partnerships like that of Union College and IBM have become almost commonplace in recent years. In 2007, Butler University received more than $10 million in donations from two private entities, the Richard L. Fairbanks Foundation and businessman (and Butler alum) Frank Levinson. The donations were used to improve Butler’s science program and upgrade the school’s supercomputer, Big Dawg, to perform 8 times faster. In October 2011, Berkshire Community College in Massachusetts launched an advanced computing program after it was awarded more than $330,000 in grant money from local research company, Nuclea Biotechnologies and the Mass. Life Sciences Center, a state-funded institute.

Other partnerships are forged exclusively between higher learning institutions. One example is the proposed Massachusetts Green High Performance Computing Center, a supercomputer network that would link Northeastern University, Harvard, Boston University, MIT and University of Massachusetts. Slated for inception in 2013, officials at these schools believe the system will have applications in virtually every college-level discipline.

In addition to private sector investors, government agencies have also made donations to colleges and/or college students for purposes of technological advancement. In Spring 2012, University of Wisconsin Mechanical Engineering Associate Professor, Dan Negrut, received a $120,000 instrumentation grant from the US Army Research Office. The grant will allow Professor Negrut to augment the school’s supercomputer, which is primarily utilized by the Wisconsin Applied Computing Center. In addition, the National Defense Science and Engineering Graduate Fellowship, sponsored by the US Department of Defense, has been conferred to roughly 3,200 students since the program was launched in 1990. The fellowship is awarded to US students in 15 different scientific disciplines, and fully finances 3 years of tuition for each recipient. The program is intended to increase “the number and quality of the nation’s scientists and engineers,” according to an article by UDaily, the school newspaper of University of Delaware.

Though collegiate supercomputing is still in its early stages, the trend stands to rise significantly in the coming years. For many campuses, the only hindrance to adopting such advanced technology is the high price tag of these computing systems. Thankfully, state agencies and private donors are seeing positive results with systems they have already invested in, and are likely to collaborate with many more colleges and universities in the coming years.

Abby Ledford

GraphLab – Workshop on Big Learning

The GraphLab workshop on large scale machine learning is a meeting place for both academia and industry to discuss upcoming challenges of large scale machine learning and solution methods. GraphLab is Carnegie Mellon’s large scale machine learning framework. The workshop will include demos and tutorials showcasing the next generation of the GraphLab framework, as well as lectures and demos from the top technology companies about their applied large scale machine learning solutions.

When – July 9th, 2012

Where – San Francisco

More info –


Danny Bickson