In message <[EMAIL PROTECTED]>,=?windows-1252?Q?Martin_MOKR
EJ=8A?= writes:
>kernel, I ran the same test twice. The first results were discarded
>but the data ended-up in kernel memory so on the second execution
>the data were just written to the disk through afs (from kernel memory).
i guess i am not convinced. with a 1G cache, i saw the following:
2.4.20(ish), SMP
relax.12% /usr/bin/time dd if=/dev/zero bs=8k count=700000 of=test.file4
700000+0 records in
700000+0 records out
2.09user 208.03system 8:56.79elapsed 39%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (159major+16minor)pagefaults 0swaps
2.6.13-rc3, SMP, PREEMPT
relax% /usr/bin/time dd if=/dev/zero bs=8k count=700000 of=test.file4
700000+0 records in
700000+0 records out
1.90user 211.66system 9:10.51elapsed 38%CPU (0avgtext+0avgdata
0maxresident)k
0inputs+0outputs (0major+130minor)pagefaults 0swaps
these results are practically the same.
_______________________________________________
OpenAFS-devel mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-devel
