| R. G. Covington, "Validation of Rice Parallel Processing Testbed Applications," Ph.D. Dissertation, Electrical and Computer Engineering Department, Rice University, Dec. 1988. |
....profiling is that the instruction set of the target processor must be the same as the instruction set of the host processor. Since it is not always possible or convenient to use a host processor that has the same instruction set as the target processor, a cross profiling technique was developed [12]. In this approach, compilers for a common source language for both the target and host machines are required. The target compiler can be a cross compiler (i.e. it can run on the host machine) or a native compiler (running on the target machine) We call the assembly language outputs of the ....
....FFT (fast Fourier transform) 4) mergesort; and (5) Global Distribution Local Sort (GDLS) The first three algorithms are numerically intensive, while the latter two include little arithmetic other than simple integer comparisons. Details of these algorithms can be found in the following theses [12, 24, 25]. 14 The eigenvalue eigenvector computation (Eigen) is characterized by single packet, singlehop communication patterns, although occasionally multi hop transmissions are used. The program is broken into three components. The first, a tridiagonal reduction, splits the data into p ( the number ....
R. G. Covington, "Validation of Rice Parallel Processing Testbed Applications," Ph.D. Dissertation, Electrical and Computer Engineering Department, Rice University, Dec. 1988.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC