It is of interest to compare the use of CINT2000 to the
use of its predecessor, CINT95. Figure 3 compares the rate
of use in the last two years before it was retired (June, 2000)
to the rate of use of CINT2000 in its first two years of in-
ception. The results for CINT95 are impressive when com-
pared to CINT2000, 70% of the papers use a majority of
the benchmarks with almost half using the full suite. How-
ever, for a mature and smaller (8 compared to 12) suite, we
expect an even higher rate of use or at least some rationale
for partial use. The latter is lacking in even greater numbers
than CINT2000. Only 5 of the 36 papers that don’thavea
perfect “batting average” present any reason for not using
the full suite.