Statistical test for comparing means when I know my dataset isn't normal?

Is there a non-parametric test for comparing the means of two sets of data (n=22 for both sets)? Both sets of data failed a normality test but appear to have equal variances.