Abstract

CONLAB (CONcurrent LABoratory) is an environment for developing algorithms for parallel computer architectures and for simulating different parallel architectures. A user can experimentally verify and obtain a picture of the real performance of a parallel algorithm executing on a simulated target architecture. CONLAB gives a high-level support for expressing computations and communications in a distributed memory multicomputer (DMM) environment. A development methodology for DMM algorithms that is based on different levels of abstraction of the problem, the target architecture, and the CONLAB language itself is presented and illustrated with two examples. Simulotion results for and real experiments on the Intel iPSC/2 hypercube are presented. Because CONLAB is developed to run on uniprocessor UNIX workstations, it is an educational tool that offers interactive (simulated) parallel computing to a wide audience.

Highlights

  • The application art>a we lun·e in mind is matrix eomputations that are ba,.;ic in most,.;cientific. economic. and t>nf!ineering: applin....,.;,..J \lm· 1'J'J:1

  • It is well known that block alf!nrithms are amenable for many paraiiPI architeetures [1-.3:

  • Because CO::'\LAB is an interactive environment it frees the user from issues like compilation, linking, etc

Read more

Summary

INTRODUCTION

Especially in an innoYatiYe phase wbere differelll idea,; and prototype implementations are examined It \Yould lw de,.,ira],]p to !Je able to expre,.;,.; the computation,., in as hif!h len·! Thi,.; paper pre,.;ents the CO:\"LAB em·ironment with focus on alf!urithm de,·elopnH"nt for di,.;trilmted memory multicomputers :D~I~l). COl\LAB iCOl\cuJTt'nt LABoratory 1 is an t>m·ironment for dt>\ t'loping: al!!orithm,.; for parallel computt>r architecture,.; and for simulatinf! This mean,.; that tlw Uf'er can experimentally 'erify and obtain a f!oml pict tire of the real performance of a parallel alf!orithm t>xecutinf! Section :2 gives an introduction to the COl\LAB hmguag

THE CONLAB ENVIRONMENT
Where Does CONLAB Run?
Extensions for Parallelism
Simulation of Parallel Execution
The Time Model for Simulation of DMM Architectures
Performance Measuring
Animated Replay of Messages
ParaGraph
ALGORITHM DEVELOPMENT METHODOLOGY
TWO EXAMPLES OF ALGORITHM DEVELOPMENT FOR DMM ARCHITECTURES
Block Torus Matrix Multiplication
Ring-Oriented Block QR Factorization
CONCLUSIONS AND FUTURE DEVELOPMENT
Results to the host
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call