| Project/Area Number |
61850063
|
|---|
| Research Category |
Grant-in-Aid for Developmental Scientific Research
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
計算機工学
|
|---|
| Research Institution | Osaka University |
|---|
Principal Investigator |
HIROAKI TERADA Professor, Faculty of Engineering, Osaka University, 工学部, 教授 (80028985)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
西川 博昭 大阪大学, 工学部, 助手 (60180593)
笹尾 勤 大阪大学, 工学部, 助手 (20112013)
浅田 勝彦 大阪大学, 工学部, 助教授 (10029093)
HIKOAKI NI@sHIKAWA Resrarch Associate, Faculty of Engineering, Osaka University
KATSUHIKO ASADA Associate Professor, Faculty of Engineering, Osaka University
TSUTOMU SASAo Research Associate, Faculty of Engineering, Osaka University
|
|---|
| Project Period (FY) |
1986 – 1987
|
| Project Status |
Completed (Fiscal Year 1987)
|
| Budget Amount *help |
¥5,000,000 (Direct Cost: ¥5,000,000)
Fiscal Year 1987: ¥2,000,000 (Direct Cost: ¥2,000,000)
Fiscal Year 1986: ¥3,000,000 (Direct Cost: ¥3,000,000)
|
|---|
| Keywords | ULSI / Highly-parallel Processing / Data-driven / Diagrammatical Language / "Flow-Thru processing" Mechanism / Pico-program / 連想処理 / トークンフローモデル |
|---|
| Research Abstract |
In designing new ULSI-orientedf processor archnitecture, it is essential to adopt completely distributed control. It is also important to have a versatile programming envionment that allows us to exploit highly-parallel processing capabilities offered by ULSI terchnology.
In orser to achieve these goals, this developmental scientific research was started in 1986 to study ULSI-oriented data-driven highly-parallel processing scheme with special emphasis on the followings.
1. to establish diadrammatical languages with much improved readbility and balidation capability.
2. to seatch for the best hardware structure to implement the language on a single ULSI chip as well as on multi-chip systems. In 1987, the research was completed successfully with the following results.
(2) It was proben by the experimental study that the use of siagrammatical languages for the pico-programmable highly-parallel processing scheme was quite fasible to run dynatically concurrent processes. (See references No.1, 2 and 4)
(2). It was also confirmed that a pico-programmable highly-parallel processing mechanism organizaed in a completely autonomous and distributed fashion will show an excellent execution rate in excess of several tens of mega-operation per second(MOPS). (See reference No.2, 3 and 5)
|
