ABSTRACT
The widespread use of reduced-instruction-set computers has generated a lot of interest in the tradeoff between the density of an instruction set and the size of the instruction cache. In this paper we present and justify a method that predicts the cache performance for a wide range of architectures, based on the miss rate for a single architecture. When we apply the method to a number of cache organizations we find that changes in code density can have a dramatic impact on memory traffic, but that modest improvements in code density do not reduce program execution time significantly in a well-balanced system.
- 1.Agarwal, A., Horowitz, M., and Hennessy, J. An Analytical Cache Model. Tech. Rept. CSL 86-304, Stanford University, September, 1986.Google Scholar
- 2.Agarwal, A. Analysis of Cache Performance for Operating System and Multi-tasking Workloads. Ph.D. Th., Stanford University, May 1987. Google ScholarDigital Library
- 3.Chow. P. and Horowitz, M. Architectural Tradeoffs in the Design of MIPS-X. 14th Annual International Symposium on Computer Architecture, IEEE, June, 1987, pp. 300-308. Google ScholarDigital Library
- 4.Davidson, J., and Vaughan, R. The Effect of Instruction Set Complexity on Program Size and Memory Performance. Second International Conference on Architectnral Support for PrOgrammin g Languages and Operating Systems, ACM/IEEE, Palo Alto, October, 1987. pp. 60-64. Google ScholarCross Ref
- 5.Ditzel. D. McLellan, H. and Berenbaum. A. Design Tradeoffs to Support the C Programming Language in the CRISP Microprocessor. Second International Conference on Architectural Support for Programming Languages and Operating Systems, ACM/IEEE, Palo Alto. October. 1987, pp. 158-163. Google ScholarCross Ref
- 6.Flynn, M. Mitchell, C., and Mulder. H. "And now a Case for More Complex Instruction Sets". IEEE Computer 20.9 (September 1987), 71-83. Google ScholarDigital Library
- 7.Gross, T. Hermessy, J., Przybylski. S. and Rowen, C. "Measurement and Evaluation of the MIPS architecture and Processor". ACM Transactions on Computer System 6.3 (August 1988). 229-257. Google ScholarDigital Library
- 8.Mitchell, C. Processor Architecture and Cache Performance. Ph.D. Th. Stanford University, June 1986. Stanford Technical Report No. CSL-86-296. Google ScholarDigital Library
- 9.Mitchell, C. Architecture and Simulation results for Individual Benchmarks. Tech. Rept. CSL-TN-86-289, Stanford University, December, 1986.Google Scholar
- 10.Patterson, D. and Sequin. C. A Reduced Instruction Set VLSI Computer. 8th Annual International Symposium on Computer Architecture, IEEE, Minneapolis, May, 1981. pp. 443-457. Google ScholarDigital Library
- 11.A. Smith. "Cache Memories". Computing Surveys 14.3 (Septembei 1982). 473-530. Google ScholarDigital Library
- 12.Steenkiste. P. LISP on a Reduced-instruction-Set Processor: Characterization and Optimization. Ph.D. Th., Stanford University, March 1987. Google ScholarDigital Library
- 13.Taylor, G.S. Hillfinger, P.N. Lams, J. et al. Evaluation of the SPUR Lisp Architecture. 13th Annual International Symposium on Computer Architecture, IEEE, Tokyo, June, 1986, pp. 444-452. Google ScholarDigital Library
Index Terms
- The impact of code density on instruction cache performance
Recommendations
The impact of code density on instruction cache performance
Special Issue: Proceedings of the 16th annual international symposium on Computer ArchitectureThe widespread use of reduced-instruction-set computers has generated a lot of interest in the tradeoff between the density of an instruction set and the size of the instruction cache. In this paper we present and justify a method that predicts the ...
A Performance Study of Instruction Cache Prefetching Methods
Prefetching methods for instruction caches are studied via trace-driven simulation. The two primary methods are "fall-through" prefetch (sometimes referred to as "one block lookahead") and "target" prefetch. Fall-through prefetches are for sequential ...
Buffering databse operations for enhanced instruction cache performance
SIGMOD '04: Proceedings of the 2004 ACM SIGMOD international conference on Management of dataAs more and more query processing work can be done in main memory access is becoming a significant cost component of database operations. Recent database research has shown that most of the memory stalls are due to second-level cache data misses and ...
Comments