Processor for performing group floating-point operations

摘要

A system and method expands a source operand to a width greater than that of a general purpose register or a data path. Operands are provided substantially larger than the data path width of a processor. The general purpose register specifies a memory address from which several data path widths of data are read. A data path functional unit is augmented with dedicated storage to which the memory operand is copied on initial execution of the instruction. Further instructions specifying the same memory address read the dedicated storage to obtain the operand value, upon verification that the memory operand has not been altered by intervening instructions. If the memory operand remains current, the memory operand fetch is combined with register operands in the functional unit, producing a result the size of a general register, so no dedicated storage is required for the result.