🚀 go-pugleaf

Options to "ls" have a great affect upon "str*" usages.  The following
results were obtained on an Amdahl 580 running uts 5.2.5:

./ls /bin /usr/bin [listed 64 times]
 %Time Seconds Cumsecs  #Calls   msec/call  Name
  22.1   10.72   10.72   80190      0.1336  strcmp
   3.3    1.60   40.25   21577      0.0742  memcpy
   0.7    0.33   47.65   10688      0.0312  strlen
   0.3    0.13   47.97                      strchr
   0.1    0.05   48.40                      memchr
   0.1    0.03   48.53                      strcpy

./ls -CF /bin /usr/bin [listed 64 times]
 %Time Seconds Cumsecs  #Calls   msec/call  Name
   7.9   10.67  100.55   80191      0.1330  strcmp
   1.2    1.58  128.38   32234      0.0491  memcpy
   0.3    0.42  133.08   21440      0.0194  strlen
   0.1    0.12  134.05       2     58.      strchr
   0.1    0.08  134.23                      memchr
   0.0    0.02  134.50                      strcpy

./ls -l /bin /usr/bin [listed 8 times]
 %Time Seconds Cumsecs  #Calls   msec/call  Name
   0.4    1.42  338.83   10023      0.1413  strcmp
   0.3    1.18  341.32   18761      0.0631  memcpy
   0.0    0.15  345.45    1336      0.112   strlen
   0.0    0.10  345.92                      memchr
   0.0    0.08  346.18                      strcpy
   0.0    0.03  346.35       1     33.      strchr

Note: on the Amdahl, strcmp and friends are in assembler.

Rick Richardson, PC Research, Inc: (201) 922-1134  ..!ihnp4!castor!pcrat!rick
	         when at AT&T-CPL: (201) 834-1378  ..!ihnp4!castor!pollux!rer