1: \begin{abstract}

2: We consider a family of unadjusted HMC samplers, which includes standard position HMC samplers and discretizations of the underdamped Langevin process. A detailed analysis and optimization of the parameters is conducted in the Gaussian case. Then, a stochastic gradient version of the samplers is considered, for which dimension-free convergence rates are established for log-concave smooth targets, gathering  in a unified framework previous results on both processes. Both results indicate that partial refreshments of the velocity are more efficient than standard full refreshments.

3: \end{abstract}

4: