Aneesh Kumar K.V wrote: > On Thu, May 15, 2008 at 11:14:51AM -0500, Eric Sandeen wrote: >> Aneesh Kumar K.V wrote: >>> On Thu, May 15, 2008 at 09:23:59PM +0530, Aneesh Kumar K.V wrote: >>>> This adds a per inode meta-block prealloc space from which >>>> meta-data block requests are served. This help in making >>>> sure meta-data block are closer. This is needed to speedup >>>> unlink of the file. Any new prealloc space is allocated near >>>> the goal block specified. The goal block is the last block >>>> allocated for the file. So we don't keep the data-block and >>>> meta-data block far apart. >>>> >>> The result can be found at >>> >>> http://www.radian.org/~kvaneesh/ext4/meta-group/ >> Out of curiosity, do you have graphs similar to >> http://www.radian.org/~kvaneesh/ext4/meta-group/ext4-noextents-metagroup.png >> >> before and after your change? > > http://www.radian.org/~kvaneesh/ext4/meta-group/exp1/ Thanks! On: http://www.radian.org/~kvaneesh/ext4/meta-group/exp1/ext4-noextents-withoutpatch-rm.png it seems interesting; do you know what's going on between ~40s and 100s? Also: between: http://www.radian.org/~kvaneesh/ext4/meta-group/exp1/ext4-noextents-read.png and http://www.radian.org/~kvaneesh/ext4/meta-group/exp1/ext4-noextents-withoutpatch-read.png it looks like about a 10% slowdown on the read.... -Eric -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html