homework代写 | mpi | 代做Algorithm | assignment – MPI Cannon’s Algorithm

MPI Cannon’s Algorithm

homework代写 | mpi | 代做Algorithm | assignment – 这是一个mpi Algorithm的practice, 考察mpi并行的理解, 涵盖了mpi/Algorithm等方面, 这个项目是assignment代写的代写题目

ass代做 assignment代写 代写assignment

2021/4/14 05 mpi Cannon’s Algorithm

05 MPI Cannon’s Algorithm

 tar,gz,tar.gzzip  4  14  0:00  4  30  23:59 17 

Implement Cannon's Matrix Multiplication  Algorithm on your own using MPI. Your implementation should
be able to handle a matrix size of 1000000x1000000 on 100 nodes.
Each node should only execute serial code, no threading. Your code should be well organized such that
each phase is broken up into well written subroutines.
Default matrix size for code performance evaluation will be provided at a later time.
Your  homework submission should be a zip file containing the following file structure and names:
  • hw5_.pdf // your writeup
  • task1 // A folder containing your solution to hw1 task 1
  • hw5_.cpp
  • hw5_.pbs
Please do not include any other files in your submission.
Evaluation and Expectations:
Your submission will be evaluated on two categories: Code Performance and Documentation.
For Code Performance, your libraries will be evaluated for both function and speed. The Rank of the
speed will determine the score earned in the code performance portion of the grade.
For Documentation, your write-up will be evaluated for how well you describe your effort and present
your results. It will also be evaluated for technical accuracy. Use the "How to Write Technical Papers" as
a guide. You do not need to use all sections of that document.
See the Homework Generic Rubric for additional guidance on how your  assignment submission will be
evaluated.

2021/4/14 05 MPI Cannon’s Algorithm

https://psu.instructure.com/courses/2109084/assignments/12710178 2 / 2

Reminder, no credit will be earned for submissions that do not have your name in the documents (Write-
up AND source code).