Machine Characterization Based on an Abstract High-level Language Machine

A linear performance model decomposes program execution into n distinct operations, such as adds, multiplies, loads, and stores. A program is characterized as an n-vector of operation counts, and a machine is characterized as an n-vector of operation times. The dot-product of the two is the time required for the machine to execute the program. A linear performance model has several uses:1. Once every machine and program is characterized, the performance of each program on each machine can be predicted without having to run them.2. Two machines (or programs) can be compared by comparing corresponding elements of their parameter vectors. The influence of individual parameters on overall performance can be used to predict the effect of design changes.3. Machines (and programs) can be classified by the similarity of their parameter vectors.

[1]  Richard A. Volz,et al.  Toward real-time performance benchmarks for Ada , 1986, CACM.

[2]  Mark Scott Johnson,et al.  Effectiveness of a machine-level, global optimizer , 1986, SIGPLAN '86.

[3]  Reinhold Weicker,et al.  Dhrystone: a synthetic systems programming benchmark , 1984, CACM.

[4]  Douglas W. Clark.  Pipelining and performance in the VAX 8800 processor , 1987, ASPLOS 1987.

[5]  Douglas W. Clark,et al.  Measuring VAX 8800 performance with a histogram hardware monitor , 1988, [1988] The 15th Annual International Symposium on Computer Architecture. Conference Proceedings.

[6]  Andrew S. Tanenbaum,et al.  Using Peephole Optimization on Intermediate Code , 1982, TOPL.

[7]  Brian A. Wichmann,et al.  A Synthetic Benchmark , 1976, Comput. J..

[8]  J. H. Griffin,et al.  Los Alamos National Laboratory computer benchmarking, 1983 , 1984 .

[9]  R. N. Ibbett,et al.  Architecture of High Performance Computers , 1989, Macmillan Computer Science Series.

[10]  Jack J. Dongarra,et al.  Performance of various computers using standard linear equations software in a Fortran environment , 1987, SGNM.

[11]  Jack J. Dongarra,et al.  Vectorizing compilers: a test suite and results , 1988, Proceedings. SUPERCOMPUTING '88.

[12]  Henri E. Bal,et al.  Language- and Machine-Independent Global Optimization on Intermediate Code , 1986, Comput. Lang..

[13]  Rafael H. Saavedra-Barrera,et al.  Machine Characterization and Benchmark Performance Prediction , 1988 .

[14]  Donald E. Knuth,et al.  An empirical study of FORTRAN programs , 1971, Softw. Pract. Exp..

[15]  Fred C. Chow,et al.  A portable machine-independent global optimizer--design and measurements , 1984 .

[16]  Hermann Luttermann,et al.  Experiences in benchmarking the three supercomputers CRAY-1M, CRAY-X/MP, FUJITSU VP-200 compared with the CYBER 76 , 1988, Parallel Comput..

[17]  Jack Dongarra,et al.  Computer benchmarking: paths and pitfalls , 1987 .