| Project/Area Number |
60550254
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
計算機工学
|
|---|
| Research Institution | Nagoya Institute of Technology (1987)
Gunma University (1985-1986) |
|---|
Principal Investigator |
SOWA Masahiro Professor of Nagoya Institute of Technology, 工学部電気情報工学科, 教授 (00008567)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
OHKUBO Motoaki Professor of Gunma University, 工学部情報工学科, 教授 (00008418)
|
|---|
| Project Period (FY) |
1985 – 1987
|
| Project Status |
Completed (Fiscal Year 1987)
|
| Budget Amount *help |
¥2,100,000 (Direct Cost: ¥2,100,000)
Fiscal Year 1987: ¥100,000 (Direct Cost: ¥100,000)
Fiscal Year 1986: ¥300,000 (Direct Cost: ¥300,000)
Fiscal Year 1985: ¥1,700,000 (Direct Cost: ¥1,700,000)
|
|---|
| Keywords | Dataflow Computer / Neumann Processing / 並列処理 / 第5世代コンピュータ / プログラムカウンタ / 遂次処理 |
|---|
| Research Abstract |
The aim of this project is to establish the method implementing of the Neumann processing in the dataflow comoputer.To implement it.it is necessary to establish the basic method for transmitting the data and control from dataflow processing to Neumann one or viceversa. This is achieved by deviding the data tokenof the dataflow processing into data and control, then put the former into the register and the latter into the program counter. The implementation is done inthe dataflow computer DFNDR-l built in Gunma University. The Neumann processing mechanism is also implemented into the high level parallel language P, which is developed by computers.
The proposed system will be a effective for the cases where dataflow computers contain a large number of serial processing, which is bottle-necks, or the data-flow computers contain a structure of data, which is not suited to the data-flow computers. By this implementation, it also becomes possible to execute the concurrent tremendous amount of softwares by the computers. This will realeze a smooth transition from serial to parallel processing.
|
