We analyze the statistics and star formation rate obtained in high-resolution numerical experiments of forced supersonic turbulence, and compare with observations. We concentrate on a systematic comparison of solenoidal (divergence-free) and compressive (curl-free) forcing (Federrath et al. 2009 a, b), which are two limiting cases of turbulence driving. Our results show that for the same RMS Mach number, compressive forcing produces a three times larger standard deviation of the density probability distribution. When self-gravity is included in the models, the star formation rate is more than one order of magnitude higher for compressive forcing than for solenoidal forcing.
