Single instruction stream parallelism is greater than two
- 1 April 1991
- journal article
- Published by Association for Computing Machinery (ACM) in ACM SIGARCH Computer Architecture News
- Vol. 19 (3) , 276-286
- https://doi.org/10.1145/115953.115980
Abstract
Recent studies have concluded that little parallelism (less than two operations per cycle) is available in single instruction streams. Since the amount of available parallelism should influence the design of the processor, it is important to verify how much parallelism really exists. In this study we model the execution of the SPEC benchmarks under differing resource constraints. We repeat the work of the previous researchers, and show that under the hardware resource constraints they imposed, we get similar results. On the other hand, when all constraints are removed except those ~equired by the semantics oft he program, we have found degrees of parallelism in excess of 17 instructions per cycle. Finally, and perhaps most important for exploiting single instruction stream parallelism now, we show that if the hardware is properly balanced, one can sustain from 2.0 to 5.8 instructions per cycle on a processor that is reasonable to design today.Keywords
This publication has 14 references indexed in Scilit:
- Available instruction-level parallelism for superscalar and superpipelined machinesPublished by Association for Computing Machinery (ACM) ,1989
- The nonuniform distribution of instruction-level and machine parallelism and its effect on performanceIEEE Transactions on Computers, 1989
- A VLIW architecture for a trace scheduling compilerIEEE Transactions on Computers, 1988
- Checkpoint Repair for High-Performance Out-of-Order Execution MachinesIEEE Transactions on Computers, 1987
- Instruction issue logic for high-performance, interruptable pipelined processorsPublished by Association for Computing Machinery (ACM) ,1987
- The hardware architecture of the CRISP microprocessorPublished by Association for Computing Machinery (ACM) ,1987
- HPS, a new microarchitecture: rationale and introductionPublished by Association for Computing Machinery (ACM) ,1985
- Measuring the Parallelism Available for Very Long Instruction Word ArchitecturesIEEE Transactions on Computers, 1984
- The Inhibition of Potential Parallelism by Conditional JumpsIEEE Transactions on Computers, 1972
- On the Number of Operations Simultaneously Executable in Fortran-Like Programs and Their Resulting SpeedupIEEE Transactions on Computers, 1972