Has the process of formation of sand roses been simulated in laboratory conditions, or does it take too long?
Yes, it has.
Cody & Cody (1998), Journal of Sedimentary Research.
Gypsum crystals were grown in experimental conditions analogous to saline terrestrial environments within bentonite clay gels by diffusion control at three different temperatures, four brine salinities, and four tannic acid (a model terrestrial humic substance) concentrations. The resulting crystals correspond to natural gypsum formed in terrestrial environments. Prismatic gypsum typically grew at both high and low temperatures in the absence of the organic additive. With increasing organic acid concentrations, the prismatic crystals progressively became flattened perpendicular to , and two temperature-dependent trends developed. At low temperatures, a hemi-bipyramidal habit dominated by 1111 faces developed, whereas the lenticular e103 dominated habit forme at higher temperatures. With progressively greater concentrations of organic material a(100) penetration twinning developed, secondary complex nucleation occurred near the twin interfaces, and finally, rosette and rosette-like aggregates formed. Higher temperatures generally favored better-formed and larger rosettes. The presence of 5% and 15% NaCl greatly decreased nucleation density and resulted in larger single crystals and crystal aggregates. The a(100) penetation twins appear to be diagnostic of gypsum growth in natural terrestrial sediments at a pH greater than 7.5.
Relevant bit about gypsum roses is in bold.