220 25845 article Path: news.gmane.org!not-for-mail From: Mathias Gaunard Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: Re: Fixed-size views and spans Date: Sun, 8 May 2016 11:16:12 +0100 Lines: 97 Approved: news@gmane.org Message-ID: References: <09bfdbe7-70d9-4d25-9f51-4b79cecfb79f@isocpp.org> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: plane.gmane.org Mime-Version: 1.0 Content-Type: multipart/alternative; boundary=001a1133b5503aa07b053251fb21 X-Trace: ger.gmane.org 1462703057 17952 80.91.229.3 (8 May 2016 10:24:17 GMT) X-Complaints-To: usenet@ger.gmane.org NNTP-Posting-Date: Sun, 8 May 2016 10:24:17 +0000 (UTC) To: std-proposals@isocpp.org Original-X-From: std-proposals+bncBDCN3ZE6W4GRBPVHXS4QKGQE62YTV2I@isocpp.org Sun May 08 12:24:00 2016 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-wm0-f72.google.com ([74.125.82.72]) by plane.gmane.org with esmtp (Exim 4.69) (envelope-from ) id 1azLsW-0000jZ-1d for gclcip-std-proposals@m.gmane.org; Sun, 08 May 2016 12:24:00 +0200 Original-Received: by mail-wm0-f72.google.com with SMTP id e201sf76414308wme.1 for ; Sun, 08 May 2016 03:23:59 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=isocpp-org.20150623.gappssmtp.com; s=20150623; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :x-original-sender:x-original-authentication-results:reply-to :precedence:mailing-list:list-id:x-spam-checked-in-group:list-post :list-help:list-archive:list-subscribe:list-unsubscribe; bh=IZWaJEU493JYDLKE0Umx9ZcScY2tBZysl8zMcPRP+ko=; b=KzS//q63ENVDcsg+tJSneXfhAWXVYIr+hL3aZmUUVXSF6y6EuFoJy/ZBLH5q/dNuju +axwZbTGKdCRyJTN/uDB2KF8OCZagnVY1QfftLI2l3Dzr9mV7BvSpjaafAzTlwXBOvER DXxhzUe8dZriGZAMkpHvgyiwjkno+D9h5cODfPhB4Pt30E1ZHSe0vHR+L0uWAWmZa8Xx 3LoI4G6gOuO9N4NZr8FpmTWo9OWyqFpfmz/J11+asC4jAIaebut8zJCC1Td9iVCl2BbQ wmG5e0sxg8witTWJp/kMTz5e62cRH8hpI8ymv0jlO1kBgM9JXcpZ8mRn7k+1W6tq997r hnAA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:mime-version:in-reply-to:references:date :message-id:subject:from:to:x-original-sender :x-original-authentication-results:reply-to:precedence:mailing-list :list-id:x-spam-checked-in-group:list-post:list-help:list-archive :list-subscribe:list-unsubscribe; bh=IZWaJEU493JYDLKE0Umx9ZcScY2tBZysl8zMcPRP+ko=; b=YiQvASG8JjhduXImaBwSH5uvp0DWiQfTq3a1gvKvNLWGWyfZB65Hs3rHDzapCdfp3V WkcEyQJxrqrGOLHC0QAwm0MUJpYuUsTODUVnXG8mWu9J658nXaEjmtkzaisJBNgMcN9A ER0qochvrRhfCgAvtUyKmyLWU0sQa5g9zO5qoSoRav53q3SRgLkEJqu8DDwWezXO/bdh KkPob+AEu/+j4Y4B1C38DYuEYLgBnC8shE2IKv0Y3lH4Hd18Pnwn8NzLfpPpEnvq6mM5 /r4IVxVd0GO/aUqQvzbdN0dDUsjINvuIyJV8/6nbqvex+UILoJCUZef6OqyTAlSYLfll 8o+w== X-Gm-Message-State: AOPr4FWCC4TeNQY4d3FySU174vgtq1RAJj4yhw8ktnt0iZXwzco61QefmyY3mZg2bVKEcQ== X-Received: by 10.28.184.199 with SMTP id i190mr503079wmf.6.1462703039168; Sun, 08 May 2016 03:23:59 -0700 (PDT) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.28.152.213 with SMTP id a204ls459584wme.27.gmail; Sun, 08 May 2016 03:23:58 -0700 (PDT) X-Received: by 10.28.10.196 with SMTP id 187mr6046941wmk.76.1462703038183; Sun, 08 May 2016 03:23:58 -0700 (PDT) Original-Received: from 7.mo2.mail-out.ovh.net (7.mo2.mail-out.ovh.net. [188.165.48.182]) by mx.google.com with ESMTPS id lc6si27971691wjc.172.2016.05.08.03.23.58 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Sun, 08 May 2016 03:23:58 -0700 (PDT) Received-SPF: pass (google.com: domain of mathias@gaunard.com designates 188.165.48.182 as permitted sender) client-ip=188.165.48.182; Original-Received: from player169.ha.ovh.net (b9.ovh.net [213.186.33.59]) by mo2.mail-out.ovh.net (Postfix) with ESMTP id D6258FFC114 for ; Sun, 8 May 2016 12:16:12 +0200 (CEST) Original-Received: from mail-lf0-f42.google.com (mail-lf0-f42.google.com [209.85.215.42]) (Authenticated sender: mathias@gaunard.com) by player169.ha.ovh.net (Postfix) with ESMTPSA id B3FE1580086 for ; Sun, 8 May 2016 12:16:12 +0200 (CEST) Original-Received: by mail-lf0-f42.google.com with SMTP id j8so172448411lfd.2 for ; Sun, 08 May 2016 03:16:12 -0700 (PDT) X-Received: by 10.112.30.174 with SMTP id t14mr5261790lbh.128.1462702572213; Sun, 08 May 2016 03:16:12 -0700 (PDT) Original-Received: by 10.25.25.85 with HTTP; Sun, 8 May 2016 03:16:12 -0700 (PDT) In-Reply-To: <09bfdbe7-70d9-4d25-9f51-4b79cecfb79f@isocpp.org> X-Gmail-Original-Message-ID: X-Ovh-Tracer-Id: 1291407197378950783 X-VR-SPAMSTATE: OK X-VR-SPAMSCORE: 0 X-VR-SPAMCAUSE: gggruggvucftvghtrhhoucdtuddrfeekledruddtgddvhecutefuodetggdotefrodftvfcurfhrohhfihhlvgemucfqggfjnecuuegrihhlohhuthemuceftddtnecu X-Original-Sender: mathias@gaunard.com X-Original-Authentication-Results: mx.google.com; spf=pass (google.com: domain of mathias@gaunard.com designates 188.165.48.182 as permitted sender) smtp.mailfrom=mathias@gaunard.com Precedence: list Mailing-list: list std-proposals@isocpp.org; contact std-proposals+owners@isocpp.org List-ID: X-Spam-Checked-In-Group: std-proposals@isocpp.org X-Google-Group-Id: 399137483710 List-Post: , List-Help: , List-Archive: List-Subscribe: , List-Unsubscribe: , Xref: news.gmane.org gmane.comp.lang.c++.isocpp.proposals:25845 Archived-At: --001a1133b5503aa07b053251fb21 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable On 8 May 2016 at 10:44, Morwenn wrote: > Another quick idea: sometimes we know that we will perform an operation o= n > the next N element of an iterable, where N is known at compile-time. It > might be the case for what we could call =C2=AB bottom-up divide-and-conq= uer =C2=BB > algorithms. For example, WikiSort > starts by dividing the > collection to sort in chunks of 8 elements and sorting them before > proceeding with soe smart merging operations. We know how to speed up som= e > operations when the number of elements is small and known at compile-time= : > sorting networks are an excellent tool to sort integers when the size of > the collection to sort is known at compile-time. > > However there is no simple way to tell to the standard library's > algorithms that they will work with a small number of elements known at > compile-time. A solution would be to add span and view classes that take = an > integer template parameter for their size, and let library implementers a= dd > overloads to algorithms for such utility classes when they think it is > useful (it could still fit the Ranges TS design). > > Any thoughts about the usefulness of such fixed-size span and view classe= s? > You can have a merge sort algorithm for arbitrary sizes be defined in terms of a kernel to sort 8 elements built with a sorting network. I already plan to provide that as part of the SIMD algorithms. --=20 You received this message because you are subscribed to the Google Groups "= ISO C++ Standard - Future Proposals" group. To unsubscribe from this group and stop receiving emails from it, send an e= mail to std-proposals+unsubscribe@isocpp.org. To post to this group, send email to std-proposals@isocpp.org. To view this discussion on the web visit https://groups.google.com/a/isocpp= ..org/d/msgid/std-proposals/CALnjya8DWi2pdiFDJOoF3G4mF7UzOW8NbA%3DN8zc5OCdqt= 6NxQw%40mail.gmail.com. --001a1133b5503aa07b053251fb21 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
On 8= May 2016 at 10:44, Morwenn <morwenn29@gmail.com> wrote:
Another quick idea: sometimes we know t= hat we will perform an operation on the next N element of an iterable, wher= e N is known at compile-time. It might be the case for what we could call = =C2=AB bottom-up divide-and-conquer =C2=BB algorithms. For example, WikiSor= t starts by dividing the collection to sort in chunks of 8 elements and= sorting them before proceeding with soe smart merging operations. We know = how to speed up some operations when the number of elements is small and kn= own at compile-time: sorting networks are an excellent tool to sort integer= s when the size of the collection to sort is known at compile-time.

= However there is no simple way to tell to the standard library's algori= thms that they will work with a small number of elements known at compile-t= ime. A solution would be to add span and view classes that take an integer = template parameter for their size, and let library implementers add overloa= ds to algorithms for such utility classes when they think it is useful (it = could still fit the Ranges TS design).

Any thoughts about the useful= ness of such fixed-size span and view classes?

<= /div>You can have a merge sort algorithm for arbitrary sizes be defined in = terms of a kernel to sort 8 elements built with a sorting network.
I al= ready plan to provide that as part of the SIMD algorithms.=C2=A0

--
You received this message because you are subscribed to the Google Groups &= quot;ISO C++ Standard - Future Proposals" group.
To unsubscribe from this group and stop receiving emails from it, send an e= mail to std-proposa= ls+unsubscribe@isocpp.org.
To post to this group, send email to std-proposals@isocpp.org.
To view this discussion on the web visit h= ttps://groups.google.com/a/isocpp.org/d/msgid/std-proposals/CALnjya8DWi2pdi= FDJOoF3G4mF7UzOW8NbA%3DN8zc5OCdqt6NxQw%40mail.gmail.com.
--001a1133b5503aa07b053251fb21-- .