Previous Table of Contents Next


TABLE 11.2 Raw Data for the Code Size Comparison Study

Processor
Benchmark RISC-I Z8002 VAX-11/780 PDP-11/70 C/70

E-String Search 140 126 98 112 98
F-Bit Test 120 180 144 168 120
H-Linked List 176 141 211 299 141
K-Bit Matrix 288 374 288 374 317
I-Quick Sort 992 1,091 893 1,091 893
Ackermann(3,6) 144 302 72 86 86
Recursive Qsort 2,736 1,368 1,368 1,642 1,642
Puzzle (Subscript) 2,796 1,398 1,398 1,398 1,678
Puzzle (Pointer) 752 602 451 376 376
SED (Batch Editor) 17,720 17,720 10,632 8,860 8,860
Towers Hanoi (18) 96 240 77 96 67
Sum 25,960 23,542 15,632 14,502 14,278
Average 2,360 2,140.18 1,421.09 1,318.36 1,298

TABLE 11.3 Ratios with RISC-I as a Base for the Code Size Comparison Study
Processor
Benchmark RISC-I Z8002 VAX-11/780 PDP-11/70 C/70

E-String Search 1.00 0.90 0.70 0.80 0.70
F-Bit Test 1.00 1.50 1.20 1.40 1.00
H-Linked List 1.00 0.80 1.20 1.70 0.80
K-Bit Matrix 1.00 1.30 1.00 1.30 1.10
I-Quick Sort 1.00 1.10 0.90 1.10 0.90
Ackermann(3,6) 1.00 2.10 0.50 0.60 0.60
Recursive Qsort 1.00 0.50 0.50 0.60 0.60
Puzzle (Subscript) 1.00 0.50 0.50 0.50 0.60
Puzzle (Pointer) 1.00 0.80 0.60 0.50 0.50
SED (Batch Editor) 1.00 1.00 0.60 0.50 0.50
Towers Hanoi (18) 1.00 2.50 0.80 1.00 0.70
Sum 11.00 13.00 8.50 10.00 8.00
Average 1.00 1.18 0.77 0.91 0.73

11.2 USING AN APPROPRIATE RATIO METRIC

Another form in which ratio games are used to show better performance is by choosing a suitable performance metric which is a ratio of two different metrics. The following example illustrates such a case.


Previous Table of Contents Next

Copyright © John Wiley & Sons, Inc.