From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.io/gmane.emacs.gnus.general/42818 Path: main.gmane.org!not-for-mail From: ShengHuo ZHU Newsgroups: gmane.emacs.gnus.general Subject: Re: elp results: entering an nnimap folder Date: Wed, 30 Jan 2002 22:10:55 -0500 Sender: owner-ding@hpc.uh.edu Message-ID: <2naduvnrk0.fsf@zsh.cs.rochester.edu> References: NNTP-Posting-Host: coloc-standby.netfonds.no Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Trace: main.gmane.org 1035178011 14214 80.91.224.250 (21 Oct 2002 05:26:51 GMT) X-Complaints-To: usenet@main.gmane.org NNTP-Posting-Date: Mon, 21 Oct 2002 05:26:51 +0000 (UTC) Return-Path: Original-Received: (qmail 27068 invoked from network); 31 Jan 2002 03:13:05 -0000 Original-Received: from malifon.math.uh.edu (mail@129.7.128.13) by mastaler.com with SMTP; 31 Jan 2002 03:13:05 -0000 Original-Received: from sina.hpc.uh.edu ([129.7.128.10] ident=lists) by malifon.math.uh.edu with esmtp (Exim 3.20 #1) id 16W7dZ-0002LJ-00; Wed, 30 Jan 2002 21:11:45 -0600 Original-Received: by sina.hpc.uh.edu (TLB v0.09a (1.20 tibbs 1996/10/09 22:03:07)); Wed, 30 Jan 2002 21:11:36 -0600 (CST) Original-Received: from sclp3.sclp.com (qmailr@sclp3.sclp.com [209.196.61.66]) by sina.hpc.uh.edu (8.9.3/8.9.3) with SMTP id VAA26258 for ; Wed, 30 Jan 2002 21:11:11 -0600 (CST) Original-Received: (qmail 27007 invoked by alias); 31 Jan 2002 03:11:09 -0000 Original-Received: (qmail 27002 invoked from network); 31 Jan 2002 03:11:08 -0000 Original-Received: from roc-24-58-72-167.rochester.rr.com (HELO zsh.2y.net) (@24.58.72.167) by gnus.org with SMTP; 31 Jan 2002 03:11:08 -0000 Original-Received: (from zsh@localhost) by zsh.2y.net (8.11.6/8.11.2) id g0V3Atq08282; Wed, 30 Jan 2002 22:10:55 -0500 Original-To: ding@gnus.org X-Attribution: ZSH X-Face: #zw9`a=vHLkn+ZH8C (Wes Hardaker's message of "Wed, 30 Jan 2002 14:08:02 -0800") Original-Lines: 59 User-Agent: Gnus/5.090006 (Oort Gnus v0.06) Emacs/21.2.50 (i686-pc-linux-gnu) Precedence: list X-Majordomo: 1.94.jlt7 Xref: main.gmane.org gmane.emacs.gnus.general:42818 X-Report-Spam: http://spam.gmane.org/gmane.emacs.gnus.general:42818 Wes Hardaker writes: > As long as everyone is submitting elp results, I thought I'd throw in > mine for a nnimap group: > > > Function Name Call Count Elapsed Time Average Time > ====================================================== ========== ============ ============ > gnus-topic-select-group 1 38.757707 38.757707 > gnus-summary-prepare 1 24.688148 24.688148 > gnus-retrieve-headers 3 11.063404 3.6878013333 > gnus-select-newsgroup 1 10.598622 10.598622 > gnus-user-format-function-b 1316 6.1292800000 0.0046575075 > gnus-thread-total-score 32501 5.6191360000 0.0001728911 > gnus-cache-retrieve-headers 1 5.080807 5.080807 > gnus-agent-retrieve-headers 1 5.079149 5.079149 > gnus-sort-threads-1 602 4.6876020000 0.0077867142 > gnus-get-newsgroup-headers-xover 1 2.974734 2.974734 gnus-summary-prepare, 63.7%, including gnus-user-format-function-b, 15.8%, and gnus-thread-total-score 14.5%. If you turn off BBDB and sorting by total-score, it may speed up a lot. gnus-select-newsgroup 27.0%, including gnus-retrieve-headers 13.1% (the time returned by elp is not actual time it consumed) and gnus-get-newsgroup-headers-xover 7.7%. If the NOV of some articles are not cached by Agent, the performance depends on the speed and latency of the network. [...] > And for an nnimap group with a large number of very old messages and > many holes in between them and the current set of messages. This > particular group takes an annoyingly long time to enter: > > > Function Name Call Count Elapsed Time Average Time > ====================================================== ========== ============ ============ > gnus-topic-select-group 1 62.883003 62.883003 > gnus-select-newsgroup 1 59.215102 59.215102 > gnus-fetch-headers 1 47.143062 47.143062 > gnus-retrieve-headers 3 35.636840000 11.878946666 > gnus-get-newsgroup-headers-xover 1 29.824552 29.824552 > gnus-cache-retrieve-headers 1 17.316289 17.316289 > gnus-agent-retrieve-headers 1 17.314588 17.314588 > gnus-summary-prepare 1 3.537631 3.537631 > gnus-summary-prepare-threads 1 3.482701 3.482701 > gnus-user-format-function-b 251 1.104247 0.0043993904 [...] This result seems much different. The running time of gnus-summary-prepare, 5.6%, is cut down a lot. But the running time of gnus-select-newsgroup, 94.1%, including gnus-retrieve-headers, 27.5%, and gnus-get-newsgroup-headers-xover, 47.4% is too much. This is similar to Daniel Pittman's report. Maybe decoding headers in gnus-get-newsgroup-headers-xover takes most of the running time. ShengHuo