From owner-freebsd-questions@freebsd.org Wed Sep 28 14:56:05 2016 Return-Path: Delivered-To: freebsd-questions@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 43658C01E89 for ; Wed, 28 Sep 2016 14:56:05 +0000 (UTC) (envelope-from galtsev@kicp.uchicago.edu) Received: from cosmo.uchicago.edu (cosmo.uchicago.edu [128.135.70.90]) by mx1.freebsd.org (Postfix) with ESMTP id 031E5CD9 for ; Wed, 28 Sep 2016 14:56:04 +0000 (UTC) (envelope-from galtsev@kicp.uchicago.edu) Received: by cosmo.uchicago.edu (Postfix, from userid 48) id 0F077CB8D27; Wed, 28 Sep 2016 09:55:57 -0500 (CDT) Received: from 128.135.52.6 (SquirrelMail authenticated user valeri) by cosmo.uchicago.edu with HTTP; Wed, 28 Sep 2016 09:55:57 -0500 (CDT) Message-ID: <37046.128.135.52.6.1475074557.squirrel@cosmo.uchicago.edu> Date: Wed, 28 Sep 2016 09:55:57 -0500 (CDT) Subject: Re: i386 version in future ? From: "Valeri Galtsev" To: "Ralf Mardorf" Cc: freebsd-questions@freebsd.org Reply-To: galtsev@kicp.uchicago.edu User-Agent: SquirrelMail/1.4.8-5.el5.centos.7 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 (Normal) Importance: Normal References: <0D6BF663-5C95-4625-B412-00E14EF97986@FreeBSD.org> <35.AF.06698.BEF4AE75@dnvrco-oedge03> <20160927134158.74c11036@archlinux.localdomain> <20160928091109.7aa95c9d@moonstudio> <20160928105420.7b1d117f@moonstudio> In-Reply-To: <20160928105420.7b1d117f@moonstudio> X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 28 Sep 2016 14:56:05 -0000 On Wed, September 28, 2016 3:54 am, Ralf Mardorf via freebsd-questions wrote: > On Wed, 28 Sep 2016 00:59:09 -0700, Mehmet Erol Sanliturk wrote: >>You are comparing different CPU brands with each other . > > It's unrealistic to compare 32 bit and 64 CPUs, produced with the same level of development. Could you provide some examples? > > Actually I'm comparing two CPUs from the same family. The CPU I'm using for my desktop computer today and the last 32 bit CPU I used for my desktop computer before I switched to 64 bit. > > My old 32 bit was an Athlon single core 800, 850 or 900 MHz, I don't remember exactly, Socket A, so the consumption was between >= 42.6 W and <= 49.7. My current, newer 64 bit dual core is an Athlon BE-2350, 2100 MHz, Socket AM2, 45 W. > >>Assume that two CPUs are produced with the SAME technology . >>A 64 bit CPU will contain more circuits than a 32 bit CPU , means more power it will use because amount of power will be proportional to number of circuit components and length of connection lines . > > But they are not produced with the same technology, the 32 bit > Athlon I mentioned requires 1.70 V or 1.75 V, the 64 bit Athlon I mentioned 1.25 V, they are anyway from the same family, they were just not produced with the same level of development. This is true: the power goes mostly into charging and discharging capacitances associated with FET gates (ans stray capacitances). Energy associated with charge is (C*(U)^2)/2. If we use half of voltage this energy drops 4 times! (hence consumed power drops 4 times). This is why the highest frequency part (core) works at lower voltage than the rest of CPU. Now, there is also capacitance C in that formulae, which roughly speaking is proportional to gate area (it is more sophisticated, yet...). So, the technology goes to twice as small nanometers (which are linear dimensions of elements), then the capacitance diminishes 4 times. That is why it only makes sense to compare "the same technology". Next, the CPUs are more sophisticated than just a width of arithmetic unit which for 64 bit would be twice than for 32 bit if CPUs were trivial devices. Actually it is not quite so, thus processing some amount of 32 bit CPU work on 64 bin CPU will (considering the rest equal) not require to spend twice the energy as compared to 32 bit CPU. You will spend more energy, this is correct, but not two times more. (Very crude analogy here is: powering off unused CPU cores; somewhat parallel utilization of components of 64 bit CPU when processing 32 bit stuff probably is less crude way of saying it). Anyway, thanks, everybody for refreshing and very insightful discussion! Valeri > > The slower CPU, slower RAM, different hardware, IDE etc., made the old computer almost half as fast, so it needs to run almost double the time, to reach the same processing power. > > https://en.wikipedia.org/wiki/List_of_CPU_power_dissipation_figures#AMD_Athlon_64_X2_.2F_Athlon_X2 http://www.anandtech.com/show/557/3 > > Regards, > Ralf > _______________________________________________ > freebsd-questions@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-questions > To unsubscribe, send any mail to > "freebsd-questions-unsubscribe@freebsd.org" > ++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++