search-icon
Biography
Hugh participated in the design and implementation of the Linux Noise Detective. The Linux Noise detective is a Linux kernel module and a GUI to collect process data directly from the kernel (on multiple cluster nodes simultaneously) and analyze the data to determine the sources of system noise. He also participated in the design and the development of the XGet file transfer software. XGet scalably transfers files to nodes within a cluster by building a tree of participants and delegating serving duties to optimal slave nodes. He participated in the development of the XCPU cluster management system. XCPU keeps the state of the cluster distributed across all nodes, allowing easy configuration of hot-spare management nodes and graceful failover that doesn't require canceling the running jobs in case of head node failure.
Presentations
Workshop
Data Analytics
HPC Center Planning and Operations
Heterogeneous Systems
Scientific Computing
State of the Practice
Datacenter
monitoring
W
Workshop
Resiliency
Scientific Computing
W
Back To Top Button