From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 20732 invoked by alias); 5 Oct 2015 21:19:04 -0000 Mailing-List: contact zsh-workers-help@zsh.org; run by ezmlm Precedence: bulk X-No-Archive: yes List-Id: Zsh Workers List List-Post: List-Help: X-Seq: 36787 Received: (qmail 7531 invoked from network); 5 Oct 2015 21:19:02 -0000 X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on f.primenet.com.au X-Spam-Level: X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham autolearn_force=no version=3.4.0 X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:from:message-id:date:in-reply-to:comments :references:to:subject:mime-version:content-type; bh=/EEHVfQJW7KXct3K0SRZqKVFja3ObSddtyK+k8dPGBc=; b=kQBWcOjGf2EqMtZADIo2PA7i+nK3Xdz0c3RqWNkJKIjhkMBwRA93Vvd6nvQgA/G/XG sSAaBjWVe94lTz8SiQqJlvcK/9RgDeiR+V+OHWAJOMEqlVqQVpYdAHZEG6FOjoma7oNz BYFXNRj9JVmqhwTWlJkEy6PXeVOuPOVhpT5vryJ2Cyv/i+JHSBFLSn3tAgMhWfuR3ao7 lNR7vVyC/9suCblCjrn6LcAiGOyxDFRClUKnm+Urwobw07SAc6kapYnhPgFljhyn2bTe 9rNcPSP7rmvpiQZQcu/AB1N4eXPnDJtkdgAKEmSB4pxVL0ZqpSZbX9fOHVBrEC2CgJpo XwmA== X-Gm-Message-State: ALoCoQmjqSaCu7uNAvD6tF2q6JsgH5zb2szJ3XzjXww0d4UKqdDBLo+8hg3OWWafQ9yMP4iGjNdA X-Received: by 10.60.58.67 with SMTP id o3mr19186577oeq.48.1444079939557; Mon, 05 Oct 2015 14:18:59 -0700 (PDT) From: Bart Schaefer Message-Id: <151005141856.ZM4458@torch.brasslantern.com> Date: Mon, 5 Oct 2015 14:18:56 -0700 In-Reply-To: Comments: In reply to Sebastian Gniazdowski "Optimization of string.c" (Oct 5, 7:23pm) References: X-Mailer: OpenZMail Classic (0.9.2 24April2005) To: Sebastian Gniazdowski , zsh-workers@zsh.org Subject: Re: Optimization of string.c MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii On Oct 5, 7:23pm, Sebastian Gniazdowski wrote: } } It might be that the faster the CPU is, the lower the gain is. If } someone could repeat the tests on machine with 1.5 GHz or more it } would be of value. I ran the tests twice each, once built without the patch (and with --enable-zsh-debug) and the other using the patch (but not debug, accidentally), on a VirtualBox VM inside a 3.2GHz host. cpuinfo in the VM guest reports the same values as the host, whether that's an accurate measure or not. I may try again later with more care to be sure the compiler flags are the same. Note in the testopt2 and testopt4 cases performance was worse with the patch, which is why I decided to send this even though I messed up the configure flags. Best viewed in fixed-width. No patch w/ debug | Patch w/o debug -----------------------------------|---------------------------------- | testopt1 | testopt1 | num calls time | num calls time -----------------------------------|---------------------------------- 1) 5 2214.75 442.95 | 1) 5 1690.37 338.07 1) 5 2191.67 438.33 | 1) 5 1697.04 339.41 | -----------------------------------|---------------------------------- | testopt2 | testopt2 | num calls time | num calls time -----------------------------------|---------------------------------- 1) 5 271.99 54.40 | 1) 5 311.95 62.39 1) 5 274.84 54.97 | 1) 5 312.72 62.54 | -----------------------------------|---------------------------------- | testopt3 | testopt3 | num calls time | num calls time -----------------------------------|---------------------------------- 1) 1 13331.32 13331.32 | 1) 1 10533.46 10533.46 1) 1 13468.76 13468.76 | 1) 1 10538.59 10538.59 | -----------------------------------|---------------------------------- | testopt4 | testopt4 | num calls time | num calls time -----------------------------------|---------------------------------- 1) 1 2085.65 2085.65 | 1) 1 2442.47 2442.47 1) 1 2010.71 2010.71 | 1) 1 2523.01 2523.01 | -----------------------------------|----------------------------------