[CS] Martin Putra Candidacy Exam/Nov 3, 2025

via cs cs at mailman.cs.uchicago.edu
Thu Oct 23 11:18:32 CDT 2025


This is an announcement of Martin Putra's Candidacy Exam.
===============================================
Candidate: Martin Putra

Date: Monday, November 03, 2025

Time:  1:30pm CST

Location: JCL 354

Title: GenoX: Exploiting Thread Signals for Dynamic CPU Scaling in Genomics Workflows

Abstract: DAG-structured genomics workflows process gigabytes to terabytes of sequencing data in shared clusters, where efficient CPU allocation is a critical resource management decision. Suboptimal allocations can significantly delay workflow executions and increase computational costs. However, existing CPU allocation approaches are often insufficient for genomic workflows due to their unique challenges: input-dependent performance variability, contention-induced interference, and dynamic workflow branching. We first characterize production-level workflows for processing DNA and RNA sequencing data, identifying three critical characteristics: inflatable parallelism (thread surges up to 96x), stall heterogeneity (up to 2x slowdowns), and millisecond-scale idle bursts that make utilization unreliable. These findings confirm that effective CPU management should satisfy three requirements: 1) responsiveness to thread bursts, 2) robustness under contention, and 3) adaptiveness to workflow dynamics. Motivated by our characterization study, we developed GenoX, an automated CPU quota management system that leverages thread counts as its primary CPU scaling signal, rather than utilization.

Operating at 5ms intervals through standard Linux cgroups, GenoX responds immediately to changes in CPU demand phases while remaining stable under contention. Evaluation shows GenoX outperforms state-of-the-art baselines, as it reduces application execution time by up to 50%, eliminates 200% straggler slowdowns, and improves workflow completion by 21%, showcasing that thread-based scaling effectively addresses the CPU allocation challenges in genomics workflows.

Advisors: Bob Grossman

Committee Members: Bob Grossman, Haryadi Gunawi, In Kee Kim



More information about the cs mailing list