ISC HPC Blog
Is Qatar About to Join the Supercomputing Club?
The Qatar National Convention Center was recently the venue for the first meeting of the User Forum of Qatar’s newly commissioned computing infrastructure institute. The National Computing Infrastructure for Research (NCIR) is the latest addition to the Qatar Foundation’s collection of R&D institutes. Is NCIR about to place Qatar on the TOP500 list and join the supercomputing club?
As I pointed out in an HPCwire article last autumn (see: HPC Prospects in Qatar, 11 November 2013), given the will and the requirements, Qatar could easily become a significant HPC power. Back in November, Qatar had just announced a short list of National Research Grand Challenges. That list included water security, energy security and cybersecurity. Strong emphasis was also given to continuing Qatar’s research efforts in biomedical sciences and human health.
Two prominent players from the US HPC and computational science community had also joined the Qatar Foundation: Thomas Zacharia, formerly Deputy Director at Oak Ridge National Laboratory, as Executive Vice President for R&D (QF R&D) and Mohammad Khaleel, formerly Division Director for Computational Sciences and Mathematics at Pacific Northwest National Laboratory, as Executive Director of QF R&D’s Qatar Energy & Environment Research Institute (QEERI). Zacharia was on record as saying “…it is within our reach in five years to have a world-leading research enterprise in Qatar.” There are now about 4 ½ years remaining on that timeline – and there are no world-leading research enterprises that don’t include HPC as an integral part.
Khaleel had assumed responsibility for the Water Security and Energy Security grand challenges in his QEERI organization. Essentially all of Qatar’s fresh water comes from desalination of either seawater or brackish ground water. It is said that Qatar’s subsurface aquifers are being depleted at a rate seven times greater than the natural recharge rate. Water security would involve more energy efficient desalination techniques and aquifer recharging, among other aspects. As Qatar is committed to reducing its carbon footprint, the energy security work will include developing new photovoltaic, battery and grid management techniques to facilitate the incorporation of large amounts of solar power into Qatar’s electricity grid. If examined in greater detail, it becomes clear that all of this research will require very significant HPC capabilities.
For cybersecurity, the need for HPC is self-evident. So, more than enough requirements were certainly in place.
How do things look six months later?
As an outcome from the Doha Round of UN climate talks in late 2012, Qatar committed itself to a significant engagement in climate change research; this research activity now appears to have been added to QEERI’s agenda. As we know, climate change research is very compute intensive.
Both the formal presentations at the NCIR User Forum and the informal discussion in its breakout sessions provided additional details related to data-intensive computing and biomedical applications.
Last November, the existence of NCIR became evident to the world through the launching of a search for a permanent Executive Director (Moe Khaleel currently serves as its acting Executive Director). The position description states that NCIR’s mission includes “Deploying and managing advanced technology to meet the large scale computing and data analysis needs of demanding science and engineering problems.” It goes on to state “The Executive Director will provide overall leadership for NCRI, ensuring the Center's position as a world leader in the development and deployment of advanced cyber infrastructure.” This certainly sounds like the description of a supercomputing center to me. Also, rumor has it that the Executive Director search may be close to a conclusion.
The first of several new buildings located in Qatar’s Education City area is scheduled to be handed over to QF R&D later this year. We understand that this building, as is also the case with others to follow, is outfitted approximately 5 megawatts of power and a 1000 m2 computer room) to accommodate a significant amount of HPC equipment.
In his keynote address at the NCIR User Forum, Khaleel stated that one should “expect NCIR tools and facilities to be accessible to Qatar entities by December.”
So, without taking too much liberty in interpretation, it would appear that Qatar is planning to join the supercomputing club by the end of this year.
If club membership is defined by having a position on the TOP500 list, it remains to speculate as to the entry point that Qatar might choose. Currently, the bottom spot on the list is occupied by a computer with a capability of just under 118 teraflops (Rmax). Getting into the top 100 would require about 335 teraflops. For the top 50 or top 10, it would be 710 teraflops and 4.3 petaflops, respectively.
If besting neighboring Saudi Arabia were an objective, that could be accomplished with 533 teraflops. Being comparable with the minimum computing capability of G8 countries would take about 358 teraflops.
If sized to meet a rough estimate of the capability needed to meet the requirements of Qatar’s very ambitious research agenda, one might expect something in the range of at least a few petaflops.
So, based on all of this, my guess is that Qatar will choose to become the 29th country to join the supercomputing club and enter it in the mid to upper reaches of the top 50. What do you think? Let me know if your analysis differs from mine.