Newman answers this question as well. His paper actually shows that every finite abelian group $G$ is the torsion subgroup of homology $H_{d-1}(K)$ for a simplicial complex $K$ on $n$ vertices, where $n = O \left(log^{1/d} |G| \right)$.