Limits on multiple instruction issue
Michael D. Smith, M.G. Johnson, Mark Horowitz · ACM SIGARCH Computer Architecture News · 1989
This paper investigates the limitations on designing a processor which can sustain an execution rate of greater than one instruction per cycle on highly-optimized, non-scientific applications. We have used trace-driven simulations to determine that these applications contain enough instruction independence to sustain an instruction rate of about two instructions per cycle. In a straightforward implementation, cost considerations argue strongly against decoding more than two instructions in one cycle. Given this constraint, the efficiency in instruction fetching rather than the complexity of the execution hardware limits the concurrency attainable at the instruction level.
Cited by 132 works. doi:10.1145/68182.68209 Open-access PDF
Authors
References
- Compilers: Principles, Techniques, and Tools
- An Efficient Algorithm for Exploiting Multiple Arithmetic Units
- Branch Prediction Strategies and Branch Target Buffer Design
- Detection and Parallel Execution of Independent Instructions
- The Inhibition of Potential Parallelism by Conditional Jumps
- On the Number of Operations Simultaneously Executable in Fortran-Like Programs and Their Resulting Speedup
- Reducing the cost of branches
- Percolation of Code to Enhance Parallel Dispatching and Execution
- Look-Ahead Processors
- Measuring the Parallelism Available for Very Long Instruction Word Architectures