<html><body class="ApplePlainTextBody" style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">On 2012-05-01, at 2:50 PM, Nathan Rutman wrote:<br><blockquote type="cite">On Apr 30, 2012, at 4:01 PM, Andreas Dilger wrote:<br></blockquote><blockquote type="cite"><blockquote type="cite">I posted some of these results in a recent presentation:<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">http://storageconference.com/2012/Presentations/T01.Dilger.pdf<br></blockquote></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">Thanks Andreas for that link.<br></blockquote><blockquote type="cite">These look like great improvements certainly - but on an absolute scale: <br></blockquote><blockquote type="cite">is one minute for 100k 0-stripe files sufficient?<br></blockquote><blockquote type="cite">4.5 min for 100k 4-stripe files? <br></blockquote><blockquote type="cite">If not, is there a particular number that we think will be acceptable? Anyone have some GPFS numbers?<br></blockquote><br>To clarify one misunderstanding here - the performance shown in the graphs is not 53 seconds for 100k 0-stripe files, but rather 53 seconds for 1.6M 0-stripe files (i.e. about 30k files per second in a 1.6M entry directory). The axis on the bottom is in units of 100k files.<br><br>That isn't to say that performance couldn't get better, but it isn't as bad as what was posted originally.<br><br>Cheers, Andreas<br>--<br>Andreas Dilger Whamcloud, Inc.<br>Principal Lustre Engineer http://www.whamcloud.com/<br><br><br><br><br></body></html>