The Letter describes the implementation of a high-connectivity process
ing node by means of an embedded shared dual-port memory. The memory i
s accessed directly by two transputers in order to realise a virtual p
rocessor with a connectivity and a computing power that are twice thos
e of a single transputer.