From mboxrd@z Thu Jan 1 00:00:00 1970 Received: by 10.96.157.163 with SMTP id wn3csp439415qdb; Thu, 12 Jun 2014 17:12:49 -0700 (PDT) X-Received: by 10.60.129.131 with SMTP id nw3mr33747075oeb.56.1402618369103; Thu, 12 Jun 2014 17:12:49 -0700 (PDT) Return-Path: Received: from rubyforge.org ([50.56.192.79]) by mx.google.com with ESMTP id y1si3166828obg.27.2014.06.12.17.12.48 for ; Thu, 12 Jun 2014 17:12:49 -0700 (PDT) Received-SPF: pass (google.com: domain of sup-talk-bounces@rubyforge.org designates 50.56.192.79 as permitted sender) client-ip=50.56.192.79; Authentication-Results: mx.google.com; spf=pass (google.com: domain of sup-talk-bounces@rubyforge.org designates 50.56.192.79 as permitted sender) smtp.mail=sup-talk-bounces@rubyforge.org; dkim=neutral (body hash did not verify) header.i=@; dmarc=fail (p=NONE dis=NONE) header.from=gmail.com Received: from localhost.localdomain (localhost [127.0.0.1]) by rubyforge.org (Postfix) with ESMTP id 000702E11F; Fri, 13 Jun 2014 00:12:49 +0000 (UTC) X-Greylist: delayed 420 seconds by postgrey-1.31 at rubyforge; Fri, 13 Jun 2014 00:11:36 UTC Received: from mail-yk0-f172.google.com (mail-yk0-f172.google.com [209.85.160.172]) by rubyforge.org (Postfix) with ESMTP id 7ED472E11F for ; Fri, 13 Jun 2014 00:11:36 +0000 (UTC) Received: by mail-yk0-f172.google.com with SMTP id 142so1536757ykq.3 for ; Thu, 12 Jun 2014 17:11:34 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:from:date:message-id:subject:to:content-type; bh=WzSo+ruhbL97TBG6C1PO+WAnA83iGq1R4WnWMy2oeq4=; b=OIYMRsoBlYPg0A3LEhO6axVf0keAviFu5ZRs/Oe5lPL7LqqyV5ClzpMCwSbyp9MLq3 DuJ1T6UCS/PDq3O2Ydw0uZGEUTNAzaOiKR3V9E3pOG5ztBp2GhgVaqAp1AINU/ZgElXv wtOkrytNhCCS0LkUGiExZ9Y0vkV5nngzj0RE1nKA1szanVIqdUyMl1oaSc5ZkSV0uMoW B/p3jWBA0oCAn1dWO9MDXYR3XQ83DJhvgFP9kAD/GApUavUmS37WRs8KlNTxZP/+JTwK MDQRyfH7I1ShK+Eg9WHURSXmKgvy3kiuUJoiphDHPQl68GF/cMnG6P/S0S9fC0vMdbY+ ktKQ== X-Received: by 10.236.154.136 with SMTP id h8mr20840446yhk.105.1402617871427; Thu, 12 Jun 2014 17:04:31 -0700 (PDT) MIME-Version: 1.0 Received: by 10.170.203.135 with HTTP; Thu, 12 Jun 2014 17:04:11 -0700 (PDT) From: Aravind Narayanan Date: Thu, 12 Jun 2014 17:04:11 -0700 Message-ID: To: sup-talk@rubyforge.org Subject: [sup-talk] Tuning Ruby's GC params for faster `sup` X-BeenThere: sup-talk@rubyforge.org X-Mailman-Version: 2.1.12 Precedence: list List-Id: User & developer discussion of Sup List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: multipart/mixed; boundary="===============5017220210607918485==" Sender: sup-talk-bounces@rubyforge.org Errors-To: sup-talk-bounces@rubyforge.org --===============5017220210607918485== Content-Type: multipart/alternative; boundary=20cf303bf88efb717a04fbac6bcf --20cf303bf88efb717a04fbac6bcf Content-Type: text/plain; charset=ISO-8859-1 Hey everyone, I'm sharing this to see if any else does this, and whether or not it makes sense to include in the defaults. `sup` was kind of slow for me, and sometimes took a few seconds to open up an (albeit large) email thread. I ran `perftools.rb` on it to see what was taking up so much time, and was surprised to find that ~ 60% of wallclock time was spent in GC. So, I tuned Ruby's GC to not run so often, resulting in GC now taking only 13% of wall clock time. Anecdotally, I also noticed that threads loaded much much faster. The params I set were https://gist.github.com/rvnd/756d6a5e82d1bba9fbb9. Anyone else notice this? Are there any other tricks people use to speed up `sup`? Regards, Aravind --20cf303bf88efb717a04fbac6bcf Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable
Hey everyone,=A0

I'm sharing this t= o see if any else does this, and whether or not it makes sense to include i= n the defaults.=A0

`sup` was kind of slow for me, = and sometimes took a few seconds to open up an (albeit large) email thread.= I ran `perftools.rb` on it to see what was taking up so much time, and was= surprised to find that ~ 60% of wallclock time was spent in GC.=A0

So, I tuned Ruby's GC to not run so often, resultin= g in GC now taking only 13% of wall clock time. Anecdotally, I also noticed= that threads loaded much much faster. The params I set were=A0https://gist.github.com/r= vnd/756d6a5e82d1bba9fbb9.=A0

Anyone else notice this? Are there any other tricks peo= ple use to speed up `sup`?

Regards,
Aravind

--20cf303bf88efb717a04fbac6bcf-- --===============5017220210607918485== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ sup-talk mailing list sup-talk@rubyforge.org http://rubyforge.org/mailman/listinfo/sup-talk --===============5017220210607918485==--