Parallel Evaluation of Functional Programs
| Project/Area Number |
01550278
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
計算機工学
|
|---|
| Research Institution | University of Tokyo |
|---|
Principal Investigator |
TAKEICHI Masato University of Tokyo, Faculty of Engineering, Associate Professor, 工学部・計数工学科, 助教授 (10011165)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KANEKO Keiichi University of Tokyo, Faculty of Engineering, Research Assistant, 工学部・計数工学科, 助手 (20194904)
IWASAKI Hideya University of Tokyo, Faculty of Engineering, Research Assistant, 工学部・計数工学科, 助手 (90203372)
|
|---|
| Project Period (FY) |
1989 – 1990
|
| Project Status |
Completed (Fiscal Year 1990)
|
| Budget Amount *help |
¥2,200,000 (Direct Cost: ¥2,200,000)
Fiscal Year 1990: ¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 1989: ¥1,400,000 (Direct Cost: ¥1,400,000)
|
|---|
| Keywords | Functional Program / Parallel Evaluation |
|---|
| Research Abstract |
We have studied a basic problem in constructing parallel evaluation system which supports functional programming. It is worth noting that recent development in hardware technology makes it practical to implement the evaluator of functional programs. It is true, however, that evaluators of functional languages implemented on conventional hardware run slower than those for procedural languages. In addition to this, it is known that there is a serious problem in evaluating functional programs in a sequential fashion. This is the reason why we study the way to evaluate functional programs in parallel.
We discuss parallel functional programming from the programmer's point of view and present novel ideas on implementing functional languages for parallel machines. First of all, we show that simple annotation works effectively to control evaluation order of parallel functional programs. And we extend this idea to make an evaluator for distributed parallel computers such as transputer systems or the Intel Hypercube. We propose a construct to represent recursive environment structures on a processor network, which is an extension to the standard environment structure.
Our research concludes with successful experimentation results to support these ideas.
|
Report
(3 results)
Research Products
(11 results)
