TY - GEN
T1 - Pathlength reduction features in the PA-RISC architecture
AU - Lee, Ruby
AU - Mahon, Michael
AU - Morris, Dale
PY - 1992/2
Y1 - 1992/2
N2 - The authors describe representative pathlength reduction features of PA-RISC (reduced instruction set computer) instructions in memory accessing, functional operations, and instruction sequencing. To illustrate the multi-op instructions in PA-RISC, comparison is made with the MIPS instruction set, rather than with some hypothetical single-op RISC instructions. It is noted that, while other RISC architectures strive to enable short cycle times and single cycle instruction execution, PA-RISC also supports pathlength reduction, without impacting either the cycle time or the CPI. Frequent operations are combined into a single multi-op instruction. Subword data are also operated on in parallel, making full use of the datapath width. Such instruction level parallelism gives PA-RISC some of the advantages of a very simple VLIW (very long instruction word) architecture (with short 32-b instructions), in addition to the inherent advantages of a streamlined RISC architecture.
AB - The authors describe representative pathlength reduction features of PA-RISC (reduced instruction set computer) instructions in memory accessing, functional operations, and instruction sequencing. To illustrate the multi-op instructions in PA-RISC, comparison is made with the MIPS instruction set, rather than with some hypothetical single-op RISC instructions. It is noted that, while other RISC architectures strive to enable short cycle times and single cycle instruction execution, PA-RISC also supports pathlength reduction, without impacting either the cycle time or the CPI. Frequent operations are combined into a single multi-op instruction. Subword data are also operated on in parallel, making full use of the datapath width. Such instruction level parallelism gives PA-RISC some of the advantages of a very simple VLIW (very long instruction word) architecture (with short 32-b instructions), in addition to the inherent advantages of a streamlined RISC architecture.
UR - http://www.scopus.com/inward/record.url?scp=0026821099&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0026821099&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:0026821099
SN - 0818626550
T3 - Digest of Papers - IEEE Computer Society International Conference
SP - 129
EP - 135
BT - Digest of Papers - IEEE Computer Society International Conference
PB - Publ by IEEE
T2 - 37th Annual IEEE International Computer Conference - COMPCON SPRING '92
Y2 - 24 February 1992 through 28 February 1992
ER -