GPTune is a performance autotuner designed particularly for HPC applications that are expensive to evaluate. GPTune uses Bayesian optimization based on Gaussian Process regression and supports advanced features such as multi-task learning, transfer learning, multi-fidelity/objective tuning, and parameter sensitivity analysis.
GPTune provides a shared database (history database) that allows users to share performance data samples, so everyone can benefit from (expensive) runs of widely used high-performance computing codes. Sign up for free to access more data and use all the available features of the history database.
- Number of tuning problems (target applications): 23
- Number of registered users: 41
- Number of function evaluations: 14132