GreenMalloc: Allocator Optimisation for Industrial Workloads

TL;DR

GreenMalloc is a framework that automatically optimizes memory allocators for industrial workloads, achieving reduced heap usage with minimal runtime impact using a multi-objective search approach.

Contribution

It introduces a novel search-based method using NSGA II to efficiently configure allocator parameters and transfer optimal settings to large system simulators.

Findings

Up to 4.1% reduction in average heap usage

Achieved 0.25% reduction in heap usage across workloads

Efficient exploration of allocator parameters from execution traces

Abstract

We present GreenMalloc, a multi objective search-based framework for automatically configuring memory allocators. Our approach uses NSGA II and rand_malloc as a lightweight proxy benchmarking tool. We efficiently explore allocator parameters from execution traces and transfer the best configurations to gem5, a large system simulator, in a case study on two allocators: the GNU C/CPP compiler's glibc malloc and Google's TCMalloc. Across diverse workloads, our empirical results show up to 4.1 percantage reduction in average heap usage without loss of runtime efficiency; indeed, we get a 0.25 percantage reduction.