Second-Generation RISC Floating Point with Multiply-Add Fused

Abstract

A 440 000-transistor second-generation RISC floating-point chip is described. The pipeline latency is only two cycles, and a double-precision result is produced every cycle. System throughput and accuracy is increased by using a floating-point multiply—add-fused (MAT) unit, which carries out a double-precision accumulate D = (A X B) + C as a two-cycle pipelined execution with only one rounding error. While the cycle time (40 ns) is competitive with other CMOS RISC systems, the floating-point performance stretches to the range of bipolar RISC systems (7.4-13 MFLOPS UNPACK). © 1990 IEEE