[Libevent-users] libevent performance : any simple http server using kqueue and perf comparison of 1.4.x with 2.x

To: "libevent-users@xxxxxxxx" <libevent-users@xxxxxxxx>

Subject: [Libevent-users] libevent performance : any simple http server using kqueue and perf comparison of 1.4.x with 2.x

From: Rv Rv <rvrv7575@xxxxxxxxx>

Date: Wed, 24 Jul 2013 14:31:29 +0800 (SGT)

Delivered-to: archiver@xxxxxxxx

Delivered-to: libevent-users-outgoing@xxxxxxxx

Delivered-to: libevent-users@xxxxxxxx

Delivery-date: Wed, 24 Jul 2013 02:31:40 -0400

Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1374647489; bh=4yFSIk9Rc6Xbwy94nfF6yTU1SnnB+NVPq750Img0ef8=; h=X-YMail-OSG:Received:X-Rocket-MIMEInfo:X-Mailer:Message-ID:Date:From:Reply-To:Subject:To:MIME-Version:Content-Type; b=5pcvzWvLWRdTjJgwXmAWYttj3jF7VmAiwpkZLwYkB/39clO8fklnPs96Vx35DJbQgptsSnLBs/DG0gQFiDBZKQVn1AKJdQgUBPhKm1ZphY+K7H+WRmsg6RKetb7SvQHRkn//knZPHcE12wv/3jlVulqOlv4sdC++5PjP+EoNBFY=

Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=X-YMail-OSG:Received:X-Rocket-MIMEInfo:X-Mailer:Message-ID:Date:From:Reply-To:Subject:To:MIME-Version:Content-Type; b=0LB4ks5j2uEfcP2pF/NBAItNp9Kdosl64aeJ0lyomQ/tC3PHsFpZs0Exjok650SR7TCYvMasVE9kSlyYGPyQUPKlVjHRl+Fzshf9UQGYHBFqnbTx/Ds8xPp4DLYNcvaf804Ic9an8zj/wLZORfeAVdVqfXZRB3/wf5AFYrhGHSg=;

Reply-to: libevent-users@xxxxxxxxxxxxx

Sender: owner-libevent-users@xxxxxxxxxxxxx

I have to eval libevent performance on our platform in the company where I work. Is there a simple kqeue based simple http server (or even an echo server would do) that I can get started on and make modifications with our logic and test its performance.

Also, are there any performance quantitative comparison between libevent 2.x and libevent1.4.x versions. I do know that libevent 2.x behaves better than libevent 1.4.12 but not sure by how much? I did see this on http://google-opensource.blogspot.in/2010/01/libevent-20x-like-libevent-14x-only.html but no other details "Preliminary results around April 2009 indicated that the rewritten codebase had improved our large-payload HTTP benchmark performance by between 35 and 174 percent."