Are supercomputers worth their super price tags?
nttnttntttnttttntttttttttnttttttntttttttnttttttttntttttttn
n n n nttttttt
“Why do we have to aim for the world’s No. 1 — what’s wrong with being the world’s No. 2?”
Ever since that short question about Japan’s vaunted K computer was posed live on national television in November 2009, Japan’s policymakers have been haunted by their need to justify the ¥112 billion of taxpayers’ money the project has soaked up since 2006.
n
The questioner was the sharp-tongued Diet member Renho, and she was directing her arresting inquiry at science and technology ministry bureaucrats summoned to one of the government’s jigyō shiwake (budget-reviewing) sessions.
Renho’s question — with its clearly implied criticism of those chasing the fleeting glory of holding the world’s top spot — has had a deep impact on the direction of the nation’s supercomputer policy since.
In fact, while bureaucrats struggled to come up with a convincing justification for the massive scale of expenditure they had overseen, the government panel adjudicating on spending judged the K computer project unnecessary and called for its budget for the following fiscal year 2010 to be frozen.
That sent shock waves through the nation’s scientific community, and a group of Nobel laureates even held a press conference to protest the proposed budget freeze. In the end, however, top politicians intervened to ensure the project’s budget was restored, and so the K computer has survived.
Nonetheless, Takahiro Hayashi, director of the Office for the Promotion of Computing Science Information at the Education, Culture, Sports, Science and Technology Ministry, acknowledges that Renho’s incisive question, and the subsequent public airing of views, were turning points for Japan’s supercomputers.
“Until the budget-screening sessions, the government policy had been increasingly geared at just achieving the world’s top spot,” Hayashi said. “We have sorted out the policies since, and now we are focusing on promoting the use of the K computer and other supercomputers at academic institutions in Japan. We want to make the supercomputers more accessible to various users.”
As part of the new policy, the science and technology ministry will connect the K computer in Kobe and other high-performance computers at universities through various networks that will make them available for use by a wide range of academic and corporate scientists, Hayashi said.
As part of this major rethink, the ministry has also specified what are termed the five “strategy fields” in which it particularly wants scientists using these computers to push for innovations. Those are:
n
nLife sciences, which include drug-design and other medical researchnNew substances/new energies, which covers the development of new power sources discovered through the atomic analysis of materialsnClimate change and disasters, including simulations on the effects of global warming and detailed analyses of the potential consequences of earthquakes and tsunamisnNext-generation manufacturing in fields such as jet engines and automotive engineeringnBasic sciences, including studies of astronomy and the universenn
One thing that’s clear from these goals is that the K computer is not intended for use in the field of national security, experts were quick to note — in contrast to most other nations developing supercomputers.
Kimihiko Hirao, director of the RIKEN Advanced Institute for Computational Science (AICS) in Kobe, where the K computer is housed, argues that Japan should show the world how supercomputers can be used for “peaceful” purposes.
“In the fall of this year or early next year, an IBM machine named Sequoia will be completed and installed at the Lawrence Livermore National Laboratory in the United States, and it is expected to achieve 20 petaflops (20 quadrillion calculations per second),” Hirao said. “But it’s a very specific machine tasked mainly with the simulations of nuclear-weapons testing.
“Japan is not a military power, but we want to show, by using the world’s best-of-the-best computers, how we can advance science.”
On a more mundane level, however, one great question still hanging over such ambitions is whether Japan will be able to nurture enough scientists capable of exploiting these computers’ untapped potential.
The science and technology ministry’s Hayashi expressed hope that the networked system will help increase the number of supercomputer users in Japan from the current 1,000 to 20,000. If it doesn’t, though, the K computer — which, on top of the ¥112 billion already poured into its development, also costs ¥10 billion a year to operate and maintain — would end up being an embarrassing national white elephant of financially petaflopic proportions.
ntnttnttntntnttnttntntnttnttntnnnt
In a time of both misinformation and too much information, quality journalism is more crucial than ever.
By subscribing, you can help us get the story right.
ntSUBSCRIBE NOWnnnntnttPHOTO GALLERY (CLICK TO ENLARGE)ntntnttnt ntttttttttntttttnttttttntttttttnttttttntttttnttttnttttttn ntnnntnttKEYWORDSntntcomputers, Riken, IBM, FLOPS, K, MEXT, supercomputersnn n n tn tttt