Using MPI to accelerate matrix multiplication algorithm

Jiaye Wu

Jiaye Wu

Unknown

0 0
  • 0 Collaborators

In real life and scientific application, matrix multiplication is needed in many places.With the help of MPI, this project realizes the acceleration of matrix multiplication operation, and the speed can be increased by 2-3 orders of magnitude in large data volume. ...learn more

Project status: Published/In Market

oneAPI, HPC

Intel Technologies
oneAPI, DevCloud

Overview / Usage

With the help of MPI, this project realizes the acceleration of matrix multiplication operation, and the speed can be increased by 2-3 orders of magnitude in large data volume.

Methodology / Approach

Allocate each row of the left matrix to each slave process, and then multiply the row with each column of the right matrix in the slave process, and then save the result matrix in the buffer array. After all the slave processes have completed the calculation, the buffer arrays will be pieced together to get the final result matrix.

Technologies Used

MPI,intel devcloud

Comments (0)