MLP Aware Scheduling Techniques in Multithreaded Processors

TL;DR

This paper proposes a new OS scheduling algorithm for multithreaded processors that leverages memory-level parallelism to improve resource utilization and overall performance.

Contribution

It introduces an MLP-aware scheduling method that dynamically balances thread MLP with system resources for better performance in multithreaded multi-core processors.

Findings

Qualitative comparison shows potential performance improvements.

Highlights importance of MLP awareness in OS scheduling.

Framework for future quantitative evaluation.

Abstract

Major chip manufacturers have all introduced Multithreaded processors. These processors are used for running a variety of workloads. Efficient resource utilization is an important design aspect in such processors. Particularly, it is important to take advantage of available memory-level parallelism(MLP). In this paper I propose a MLP aware operating system (OS) scheduling algorithm for Multithreaded Multi-core processors. By observing the MLP available in each thread and by balancing it with available MLP resources in the system the OS will come up with a new schedule of threads for the next quantum that could potentially improve overall performance. We do a qualitative comparison of our solution with other hardware and software techniques. This work can be extended by doing a quantitative evaluation and by further refining the scheduling optimization.