David Wright wrote on 04/27/2015 17:29: > Quoting Jörg-Volker Peetz (jvpe...@web.de): >> Correction of the "jot" command arguments: >> >> Jörg-Volker Peetz wrote on 04/25/2015 16:15: >>> <snip> >>>> >>>> for i in `seq 5000` >>>> do >>>> date=$((10000000+i)) >>>> cat <<EOF > "$dir/cur/$date.1.host:2,S" >>> <snip> >>> >>> the situation becomes much worse if you generate the filenames from a random >>> sequence. Try to replace the command "seq" by "jot" from the package >>> athena-jot like >>> >>> for i in $(jot 5000 1 5000) >>> >> for i in $(jot -r 5000 1 5000) >> >>> That makes the numerical order of the files have a random i-node number >>> sequence . > > Apart from the fact that it doesn't (it only randomises the > filenames), it won't even generate 5000 files because of > duplication. > Have you tried benchmarking both cases? I definitely see very different timings (on ext4) for the grep benchmark. Maybe my wording is unclear. I haven't had a look how "grep -r" traverses the directory contents. I thought it would search the files in a sequence according to some lexicographical order of their names. -- Regards, jvp.
-- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/mho1ej$d61$1...@ger.gmane.org
