On Optimal Resource Allocation Algorithms for Paralellisable Jobs

On Optimal Resource Allocation Algorithms for Paralellisable Jobs

Computer Science

Aadya Mudgil

2024

Arpan Mukhopadhyay

Unprecedented demands are being placed on compute resources due to HPC jobs such as next-gen machine learning. Distributed systems rely on resource provisioning algorithms to parallelise these jobs across many nodes; however, traditional algorithms are unlikely to perform- an inefficient and costly waste of power and time. This research aims to design the optimal algorithm for parallelisable tasks, under a real-world workload, and validate theoretical expectations with the results from system simulations.

resource allocation algorithms, algorithm optimisation, optimisation, paralellisation, hpc, servers, convex optimisation, computer science, maths, performance