Authors:
Julio Corona
1
;
Pedro Rodrigues
1
;
Mário Antunes
1
;
2
and
Rui Aguiar
1
;
2
Affiliations:
1
Instituto de Telecomunicações, Universidade de Aveiro, Aveiro, Portugal
;
2
DETI, Universidade de Aveiro, Aveiro, Portugal
Keyword(s):
Machine Learning, Heterogeneous Computing, MLOps, DevOps.
Abstract:
The rapid evolution of hardware and the growing demand for Machine Learning (ML) workloads have driven the adoption of diverse accelerators, resulting in increasingly heterogeneous computing infrastructures. Efficient execution in such environments requires optimized scheduling and resource allocation strategies to mitigate inefficiencies such as resource underutilization, increased costs, and prolonged execution times. This study examines the computational demands of different stages in the Machine Learning Operations (MLOps) pipeline, focusing on the impact of varying hardware configurations characterized by differing numbers of Central Processing Unit (CPU) cores and Random Access Memory (RAM) capacities on the execution time of these stages. Our results show that the stage involving resource-intensive model tuning significantly influences overall pipeline execution time. In contrast, other stages can benefit from less resource-intensive hardware. The analysis highlights the impor
tance of smart scheduling and placement, prioritizing resource allocation for model training and tuning stages, in order to minimize bottlenecks and enhance overall pipeline efficiency.
(More)