barenya255/parallel_job_scheduling

Assignment 4 of GPU Programming

parallel_job_scheduling

Input file should be formatted as follows :

The first line contains the number of computer centres.

For each computer centre,

– computer centre number, number of facility rooms P – The next line contain P space-separated integers i.e. facility room numbers – The next line contain P space-separated integers i.e. capacity C of each facility room

The next line contain the total number of requests R.

The next R lines contains details of each request from user which are followed below,

– Request ID, computer centre number, facility room number, Starting slot number, number of slots to be reserved.

Optimizations used :

• High degree of parallelization given (almost constant runtime with respect to number of requests.)
• Graph created in CSR format corresponding to the cluster's architecture.
• Matching algorithm applied.

Using the scheduler :

• make a folder INPUT_DIR
• place all inputs inside the folder
• run make

Dynamic Version

• uses producer-consumer analogy to make almost parallel updation.
• Will be implemented shortly.