The 2.1 caching code seems to be caching a lot less than 2.0. :(
On the bright side look at the "rm" timings!!!
Y.
(Dual PPro 200, 64mb ram, Advansys 940U, IBM Ultrastar 2ES)
Linux 2.0.31 SMP
[root@miron /tmp]# time cp -rf /usr/src/linux .
0.17user 5.62system 0:43.43elapsed 13%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (13778major+72minor)pagefaults 0swaps
[root@miron /tmp]# time cp -rf /usr/src/linux .
0.16user 6.83system 1:09.34elapsed 10%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (6881major+71minor)pagefaults 0swaps
^^^^
[root@miron /tmp]# time rm -rf linux
0.04user 0.77system 0:16.35elapsed 4%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (74major+12minor)pagefaults 0swaps
Linux 2.1.72 SMP
[root@miron /tmp]# time cp -rf /usr/src/linux .
0.11user 4.36system 0:37.17elapsed 12%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (13811major+72minor)pagefaults 0swaps
[root@miron /tmp]# time cp -rf /usr/src/linux .
0.20user 5.10system 1:16.49elapsed 6%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (13896major+71minor)pagefaults 0swaps
^^^^^
[root@miron /tmp]# time rm -rf linux
0.03user 0.60system 0:06.29elapsed 10%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (74major+12minor)pagefaults 0swaps