News on FGCS Technology and related activities by Research Institute for Advanced Information Technology (AITEC), the successor of ICOT
Many of you may be wondering what happened to the contest. We are sorry for the delay in sending this AITEC NEWS and following up on the contest. Over the last few months, AITEC has embarked on various new activities. In this issue, we focus on the KLIC Programming Contest and introduce news of some other activities.
From the end of October to early November, I and two other researchers went on business to the U.S.A. to study the R&D funding systems of the U.S. government. We visited several funding agencies such as the NSF (National Science Foundation) and DARPA (Defense Advanced Research Projects Agency).
We enjoyed the hospitality of the people we met on this trip, even though we had got to know some of them only through the WWW. We arranged the visit solely through e-mail with those we met on the web and had useful meetings with them in the U.S.
As we learned at ICOT when it received nearly 300 visitors a year from other countries, it is very important to prepare well in order to receive visitors, exchange information productively and make the visit useful.
It is often said that Japan imports more information than it exports, especially in the IT field and that it hides important information behind secret veils. Although we feel that there are major language barriers and cultural gaps between Japan and overseas, we must strive to openly exchange information to keep good relationships with the rest of the world. For example, it is becoming increasingly important to utilize the Internet and to prepare information in English to distribute it to the world.
AITEC is, of course, trying to provide you with as much useful information as possible on our activities.
Let's begin with the headlines of AITEC NEWS No. 6.
(Makiko Sato)
The KLIC Programming Contest was held for the purpose of disseminating the KL1 parallel logic programming language. We believe that this is the first contest in Japan for a born parallel programming language and maybe in the world; we finally received 83 entries including one from the U.K.
The contest administrators were delighted that the number of programs received exceeded expectations, and are now examining the programs, such as measuring the execution time of the programs and evaluating the documents including manuals.
The nominees will be decided at the contest committee meeting on December 11, and personally informed by December 13.
The results of the contest will be announced publicly on December 19, on
our experimental "cyber" awards ceremony on the AITEC Home Page. Check it out!
(http://www.icot.or.jp) (We are sorry that the cyber ceremony is in
Japanese only, but planning to prepare the English version for the next
contest.)
(Article No. 6-1)
2. Report from Department of Future Technology Studies
=======================================================
AITEC consists of three departments: the department of General Affairs, the department of FGCS Technology and the department of Future Technology Studies. The department of FGCS Technology conducts activities such IFS dissemination events like the KLIC Programming Contest. The department of Future Technology Studies researches leading-edge information technology and offers the research results to help the government draw up new policies and schemes, or business leaders to make better R&D plans.
The department of FTS started this fiscal year with five new research staff members. One of the recent research projects of the department of FTS is about how to increase Japanese IT competitiveness in which we investigate why Japan has lagged behind in the field of IT compared to other countries.
We plan to investigate this problem from following two points of view: One is lack of past investment to scientific and pre-competitive researches which would produce many intellectual sources of new software technologies and also lack of domestic collaboration between academia and industry as well as international collaboration.
The other is lack of "open and competitive" systems to guide government sponsored projects which effectively seek for seeds of R&D projects and encourage researchers and industrialists to cultivate new markets. We plan to organize a committee to discuss this problem soon.
Another study is listing up and analyzing new technologies that are expected to be important in the industries in the future.
The department of FTS has started two follow-up working groups: the
"Petaflops Machine and Supercomputing Technology Research Working Group
(Chairman: Dr. Yoshinori Yamaguchi, Electrotechnical Laboratory) (PFM-WG)";
and "Network and AI-related New Technologies Research Working Group
(Chairman: Dr. Hiroshi Okuno, NTT Basic Research Laboratories) (NAI-WG).
The reports of these working groups will be given on the AITEC Home Page.
Please take a look.
(No related articles)
3. Report of IEEE Symposium "Frontiers '96" at Annapolis ========================================================= Two researchers of the department of FTS participated in "The Sixth Symposium on Massively Parallel Computation" subtitled "Frontiers '96" from October 27 to 31 at Annapolis, Maryland in the U.S.A. The theme of the symposium was petaflops computing and related topics.
As the basic technologies for teraflops computing become feasible from a research viewpoint, some researchers in the U.S. are working on petaflops machines which will be thousands of times faster than teraflops machines by the year 2014. The department of FTS is starting to study the issue of petaflops machines.
If interested in the report on "Frontiers '96," you will find the
"Frontiers '96" site here. (http://www.aero.hq.nasa.gov/hpcc/front96.html)
(No related articles)
4. Research Trip to the U.S. Research Founding Agencies ======================================================= From the end of October to early November, the department of FGCS technology sent three researchers to the U.S.A. to research how U.S. founding systems are organized and operated so effectively. They visited the Pittsburgh Supercomputing Center (PSC), Carnegie Mellon University (CMU), National Science Foundation (NSF), Defense Advanced Research Projects Agency (DARPA) and Office of Sponsored Programs (OSP) at MIT.
They found that the U.S. systems are operated on an "open and competitive" basis which is very different from the Japanese system. This difference seems to be due to the philosophical difference of establishment in each country. An analysis of the difference should prove to be very interesting for the department of FTS. (No related articles)
5. IFS Seminar at Kyushu University
====================================
AITEC holds IFS seminars at universities in Japan and overseas in response to the needs of KLIC users who continue to increase.
The most resent seminar was held at the Graduate School of Information Science and Electrical Engineering, Kyushu University on December 5 and 6. Lectures on KLIC systems with KL1 programming practice, and demonstrations of a parallel theorem prover on a 128-processing-element PIM/m were held for about 20 graduate students and teaching staff of the university.
We highly recommend you to learn KL1 parallel language because the parallel and distributed machines become more important and more popular. KL1 is designed based on the sound logic system and has less irregular rules in it. Furthermore, you can port it on most sequential UNIX Workstations and many commercial parallel machines. Please contact us, if you are interested in hosting IFS seminars. (irpr@icot.or.jp) (No related articles)
1) The submission of programs for the contest closed.
The administration office stopped accepting programs for the contest on November 17, Sunday, having had 83 entries (including group entries) in total.
At first, few programs arrived by e-mail and the staff were worried what was wrong, but by the time the deadline came they had received many more programs than expected.
On Sunday afternoon on the final day for program submission, the staff finished work for the day thinking that no more programs would arrive, and went home. But on Monday morning when they came to the office, 20 more programs had arrived.
The staff were delighted to know that the applicants, especially students, had been working on Saturday and Sunday for the contest. I guess there should be no Saturday or Sunday for students anyway!
The contest administration sent all the applicants acknowledgments and finished accepting programs.
However, they are still concerned whether they might have missed some applications due to a fault in electronic mailing systems. Please contact us if you entered but received no reply.
The number of applications in each category was as follows:
Category 1: 54 programs for the given subject in a sequential environment
Category 2: 15 programs for the given subject in a parallel environment
Category 3: 8 programs on any subject
There were eight applications for category 3. It seems few but we are satisfied with the number received because we believe that tremendous preparations must be needed in order to complete the programs.
For category 2, we were glad to receive 15 applications, because we imagine that there are few parallel machines available to most students. It is in this category that we hope to attract more applicants in the future. We look back and wonder whether we should have announced that we could provide applicants with parallel machines for debugging and tuning.
The submitted programs are now being evaluated by the contest committee members and technical staff. The final committee meeting will be held on December 11, Wednesday to decide the nominees of the awards. We look forward to seeing the results.
If you would like to know more about the contest, please look at the AITEC home page for the contest. (http://www.icot.or.jp)
2) Progress of program evaluation
The contest administrators are now evaluating the programs.
The KLIC Programming Contest Executive Committee has been examining and evaluating the programs.
For category 1, the administration staff examined the programs with a simple test program. As a result, about 10 of the programs failed the test. We are sorry and hope that the applicants who failed will participate again in the next contest.
A few applicants seem to be new to KL1 programming judging from their documents, and so we're glad that some of them were able to pass the test.
The programs which passed the first test program can go on to the second stage of the examination in which other test programs will be used with various data patterns. We will decide the finalists after evaluating the run times, quality of the source program and the documents.
Again for category 2, the committee members tested the programs in the same manner as for category 1. As a result, three programs failed the test program.
Some of the committee members commented that if a program does not become faster as the number of processors is increased, it does not deserve to win an award. But since this is the very first KLIC programming contest and some applicants had to make their programs without real parallel machines, the contest administration hopes to give as many awards as possible. We are now waiting for the committee's decision.
For category 3, there is no specific evaluation data as for two other categories. The evaluation will more or less depend on the committee members personal judgment, but the administration will be as fair as possible.
We would like to stress that the judgment of the contest is fair and impartial, with judges not knowing the applicants' names.
2. "Cyber" Awards Ceremony Coming Soon
======================================
The award nominees will be decided by the evening of December 11, and will be personally informed by December 13.
The award nominees will be publicly announced in the "cyber" ceremony on the AITEC Home Page. The chairman of the contest committee will give his impressions of the contest and also the nominees' speeches will be published at the "cyber" ceremony. We are also looking forward to seeing the atmosphere of the ceremony using a virtual-reality presentation. This will go on-line on December 19, Thursday. Don't miss it! (http://www.icot.or.jp)
Not only the participants of the contest, but also the editorial staff are excited about the results and wonder who will receive the awards. But in any case, we are pleased that many people tried to program with KL1 and got to know about KLIC.
If you could not participate in this contest, please don't worry - we look forward to your entry in the next KLIC contest next year. We also plan to welcome more foreign applicants next year.
In the next AITEC NEWS, we will announce the award nominees.
See you soon.
********************************************************************** * * * A I T E C N E W S Issue #6 * * AITEC NEWS Editorial Team: * * Makiko Sato, Chie Takahashi, Akira Aiba * * Hiroshi Sato, Shunichi Uchida * * Issued: January 16, 1997 * * By: Research Institute for Advanced Information * * Technology (AITEC), a subcenter of * * Japan Information Processing Development * * Center (JIPDEC) * * 2-3-3, Minato-ku, Shiba, Tokyo 105, Japan * * Tel: 03-3456-3191 Fax: 03-3455-4877 * * E-mail: aitec-news@icot.or.jp * * * **********************************************************************