Timothy Zhu
Pennsylvania State University, PA·Computer Science·Early-career
Computer systems waste resources in two ways: they discard hardware that still mostly works, and they run slowly for reasons engineers struggle to find. Producing new computers creates heavy carbon emissions, and tracking down slowdowns in systems built from thousands of components takes specialized teams.
- Fields
Computer Science
Engineering
- Topics
computer systems
distributed systems
performance debugging
- Grants
- 2 known grants, through Apr 2028
What they work on
- Develop software approaches for managing partially broken consumer and datacenter machines so their remaining working components can be used as-is, rather than repaired or discarded.
- Extend the effective lifetimes of solid state drives (SSDs), which wear out with use, by building age-aware scheduling and management that factors device wear into decisions.
- Create tools that automatically generate experiments to confirm or refute performance hypotheses in distributed systems, progressively narrowing down the root causes of slowdowns.
- Integrate code analysis and performance modeling into open-source diagnosis tools that find fundamental design and implementation inefficiencies, not just faulty components.
More about the research
Both projects build software that makes computing systems last longer and run more efficiently.
Why it matters
Reusing partly broken computers and stretching SSD lifetimes could reduce how much new hardware must be made. That would lower carbon emissions and toxic manufacturing byproducts. Automated performance diagnosis tools could save engineering time for software developers and companies running large distributed systems, and could help them build more cost- and energy-efficient services.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Mahmut Kandemir
Pennsylvania State University, PA·Computer Science
Characterize diverse HPC workloads on real and simulated computational storage devices to identify which computations benefit from running inside storage, and what performance tradeoffs result.
Abhinav Bhatele
University of Maryland, MD·Computer Science·Early-career
Build machine learning models of parallel application performance from collected software performance data, and use them to automatically select high-performance configurations instead of relying on hand tuning.
Yu David Liu
Binghamton University, NY·Computer Science
Develop an operating system principle that continuously measures, tracks and rewards the sustainability impact of software activity on smartphones.
The school
Pennsylvania State University: 60 labs on PI Pages, $1.3B a year on research, 743 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.