In my humble opinion; there is no such thing as a magic box with default settings in de novo assembly.

We tried the Ngen as well as one of its direct compettitors CLCbio genomic workbench...which also works from a desktop computer for smaller sized genomes....you still need quite some computational power if you want to do large eukaryotic stuff..... Currently for Illumina data 150bp+ we are quite comfortable with multiple kmer assemblies using ABYSS.

My 2ct.