Phylogenetic trees are important in biology since their applications range from determining protein function to understanding the evolution of species. Maximum Likelihood (ML) is a popular optimization criterion in phylogenetics. However, inference of phylogenies with ML is NP-hard. Recursive-Iterative-DCM3 (Rec-I-DCM3) is a divide-and-conquer framework that divides a dataset into smaller subsets (subproblems), applies an external base method to infer subtrees, merges the subtrees into a comprehensive tree, and then refines the global tree with an external global method. In this study we present a novel parallel implementation of Rec-I-DCM3 for inference of large trees with ML. Parallel-Rec-I-DCM3 uses RAxML as external base and global search method. We evaluate program performance on 6 large real-data alignments containing 500 up to 7.769 sequences. Our experiments show that P-Rec-I-DCM3 reduces inference times and improves final tree quality over sequential Rec-I-DCM3 and stand-alone RAxML.