Design and Implementation of the VAX 8600 Pipeline

VAX 8600 pipelining allows the processing of as many as four macroinstructions at once. In addition, the machine cycle time is cut by more than half. Pipelined architecture increases both the speed and complexity of the new VAX 8600. The VAX 8600 uses parallel processing to reduce the effective number of cycles required to execute instructions, in addition to speeding up machine cycle time by a factor of 2.5 over the VAX-11/780. We will assume readers are familiar with conventional VAX architecture and instructional format. This article describes the VAX 8600 architecture, relationships between the major CPU subsections, instructional format, and solutions to problems created by the complexity of the unit's pipelined design .