From mboxrd@z Thu Jan 1 00:00:00 1970 From: Joe Landman Subject: Re: crush: straw is dead, long live straw2 Date: Fri, 12 Dec 2014 13:18:57 -0500 Message-ID: <548B3191.4040605@gmail.com> References: <20141212091408.GN4150@thinkpad.thebehrens.net> <548B0C16.4000509@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mail-ie0-f180.google.com ([209.85.223.180]:53155 "EHLO mail-ie0-f180.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1031125AbaLLSTA (ORCPT ); Fri, 12 Dec 2014 13:19:00 -0500 Received: by mail-ie0-f180.google.com with SMTP id rp18so7144285iec.25 for ; Fri, 12 Dec 2014 10:18:59 -0800 (PST) In-Reply-To: Sender: ceph-devel-owner@vger.kernel.org List-ID: To: Sage Weil Cc: Thorsten Behrens , ceph-devel@vger.kernel.org On 12/12/2014 11:20 AM, Sage Weil wrote: > We can't use floating point. The code needs to run in the kernel. We > also need the rseults to be perfectly deterministic and consistent > across all architectures; I'm not sure if all floating point > implementations (and log implementations) will do that? Have a look at this (using the code I pointed to before): https://gist.github.com/joelandman/ec6f3abef9bc5f1c7b0e Running on my desktop box, library double function (with casts) is about 3x slower than the local log_2 version.