Flux, developed by researchers from Lawrence Livermore National Laboratory and University of Tennessee, Knoxville, is a next-generation workload management framework for supercomputers, high-performance computing (HPC) clusters, servers in the cloud and laptops. It combines fully hierarchical resource management with graph-based scheduling to improve the performance, portability, flexibility and manageability of scheduling and execution of complex scientific workflows on HPC systems both at the system and user level.
Flux manages massive numbers of processors, memory and other resources of a computing system and assigns the work requested by users — also known as jobs or workloads — to one or more available resources that complete the work, a method known as scheduling. A job is typically expressed in a script that contains a formal specification that requests resources, identifies applications along with their input data and environment and describes how to deliver the output data.
Workload management software like Flux is critical for HPC users because it enables efficient execution of user-level applications while simultaneously providing the HPC facility with tools to maximize overall resource utilization.