Lines Matching +full:system +full:- +full:wide

85 We prefer to run a profiling system for about a one day
93 as measured by the number of times each system call
106 in the long term system performance is negligible,
108 but the long term system performance improves significantly.
117 representing 40% of the time processing system calls,
138 The system measurements collected showed the
167 cache we ran ``ls \-l''
169 Before the per-process cache this command
170 used 22.3 seconds of system time.
172 of user time, but the system time dropped to 3.3 seconds.
174 This change prompted our rerunning a profiled system
178 still accounted for 36% of the system call time,
180 This amounted to a drop in system time from 57% to about 55%.
195 Figure 4. Call times for \fInamei\fP with per-process cache.
213 Thus Robert Elz introduced a system wide cache of most recent
225 then the per-process cache may still be useful in
232 Depending on the size of the system,
234 using 10-44 kilobytes of physical memory.
237 After adding the system wide name cache we reran ``ls \-l''
240 however the system time rose slightly to 3.7 seconds.
245 Another profiled system was created and measurements
248 \fInamei\fP accounting for only 31% of the system call time,
251 System time dropped from 55% to about 49%.
273 On the profiled system a 60% hit rate was observed in
274 the system wide cache.  This, coupled with the 25%
275 hit rate in the per-process offset cache yielded an
277 While the system wide cache reduces both the amount of time in
280 it is interesting to note that the actual percentage of system