Improvements to Supercomputing Service Availability Based on Data Analysis

As the demand for high-performance computing (HPC) resources has increased in field of computational science, an inevitable consideration is service availability large cluster systems such as supercomputers. In particular, factor that most affects supercomputing services job scheduler utilized allocating resources. Consequent to submitting user data through analysis, 25.6% jobs failed because program errors, or I/O errors. Based on this we propose a K-hook method scheduling increase success rate submissions and improve services. By applying method, job-submission was improved by 15% without negatively affecting users’ waiting time. We also achieved mean time between interrupts (MTBI) 24.3 days maintained average system at 97%. research verified Nurion supercomputer real environment, value expected be found significant improvements.