Torque
In 2006, Garrick Staples presented the Terascale Open-source Resource and Queue Manager at the ACM/IEEE conference on Supercomputing. This software emerged from the Portable Batch System, a predecessor that laid the groundwork for modern cluster management. The TORQUE community expanded capabilities to improve scalability over time. Fault tolerance became a key focus as developers sought to handle larger workloads without failure. Overall functionality grew through iterative updates driven by high-performance computing needs. Early versions focused on basic job oversight before evolving into complex distributed resource managers.
The system oversees batch jobs across multiple compute nodes in a distributed environment. Administrators gain control over scheduling tasks that run on clusters of computers. Utilization rates improve when resources are allocated efficiently to active processes. Distributed nodes communicate with the central manager to report status and request new assignments. Scheduling algorithms determine which jobs execute first based on priority and available capacity. Administration tasks include monitoring node health and managing user access permissions.
Organizations can pair TORQUE with the non-commercial Maui Cluster Scheduler for enhanced features. Commercial entities often choose the Moab Workload Manager to optimize large-scale environments. These integrations provide additional tools beyond the base resource manager's native functions. Maui offers free optimization options while Moab delivers enterprise-grade support packages. Both partners enable better handling of complex cluster architectures found in research facilities. The choice between them depends on budget constraints and specific operational requirements.
NCSA contributed significant development efforts to expand the project's reach. Sandia National Laboratories provided critical testing infrastructure during early deployment phases. TeraGrid served as a major platform where the software demonstrated its capabilities at scale. USC and OSC also added code improvements to enhance stability and performance. The US Department of Energy supported funding initiatives that accelerated adoption across universities. PNNL and UB joined these efforts to refine fault tolerance mechanisms over time.
June 2018 marked a turning point when TORQUE lost its open-source designation. Licensing conflicts arose because the software used the OpenPBS version 2.3 license. Debian Free Software Guidelines classified the program as non-free despite earlier descriptions. This shift forced many users to reconsider their reliance on the tool for public projects. Developers had to navigate new legal frameworks before continuing distribution. The controversy highlighted tensions between proprietary restrictions and community-driven development goals.
Common questions
When did Garrick Staples present the Terascale Open-source Resource and Queue Manager at the ACM/IEEE conference on Supercomputing?
Garrick Staples presented the software in 2006. This presentation occurred during the ACM/IEEE conference on Supercomputing.
What licensing conflict caused TORQUE to lose its open-source designation in June 2018?
TORQUE lost its open-source designation because it used the OpenPBS version 2.3 license. The Debian Free Software Guidelines classified the program as non-free due to this specific license type.
Which organizations contributed development efforts or funding to expand the reach of TORQUE?
NCSA, Sandia National Laboratories, USC, OSC, PNNL, and UB all contributed code improvements or testing infrastructure. The US Department of Energy provided funding initiatives that accelerated adoption across universities.
How do organizations pair TORQUE with other schedulers for enhanced features?
Organizations can pair TORQUE with the non-commercial Maui Cluster Scheduler for free optimization options. Commercial entities often choose the Moab Workload Manager to optimize large-scale environments.