From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on yquem.inria.fr X-Spam-Level: X-Spam-Status: No, score=0.0 required=5.0 tests=none autolearn=disabled version=3.1.3 X-Original-To: caml-list@yquem.inria.fr Delivered-To: caml-list@yquem.inria.fr Received: from concorde.inria.fr (concorde.inria.fr [192.93.2.39]) by yquem.inria.fr (Postfix) with ESMTP id BA65BBC69 for ; Tue, 14 Aug 2007 18:53:19 +0200 (CEST) Received: from capsaicin.mamane.lu (5.xs4all.nl [82.95.233.223]) by concorde.inria.fr (8.13.6/8.13.6) with ESMTP id l7EGrIXZ012690 (version=TLSv1/SSLv3 cipher=AES256-SHA bits=256 verify=NO) for ; Tue, 14 Aug 2007 18:53:19 +0200 Received: from master by capsaicin.mamane.lu with local (Exim 4.67) (envelope-from ) id 1IKzds-0003s2-AT; Tue, 14 Aug 2007 18:53:16 +0200 Date: Tue, 14 Aug 2007 18:53:16 +0200 From: Lionel Elie Mamane To: "Koprowski, A." Cc: Richard Jones , caml-list@yquem.inria.fr Subject: Re: [Caml-list] Utilizing more than 4GB of memory in caml? Message-ID: <20070814165316.GA14379@capsaicin.mamane.lu> Mail-Followup-To: Lionel Elie Mamane , "Koprowski, A." , Richard Jones , caml-list@yquem.inria.fr References: <20070814143423.GA29829@capsaicin.mamane.lu> <20070814161741.GA8221@capsaicin.mamane.lu> <20070814162559.GB1651@furbychan.cocan.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: X-Operating-System: GNU/Linux X-Request-PGP: http://www.mamane.lu/openpgp/rsa_v4_4096.asc User-Agent: Mutt/1.5.13 (2006-08-11) X-Miltered: at concorde with ID 46C1DDFF.000 by Joe's j-chkmail (http://j-chkmail . ensmp . fr)! X-Spam: no; 0.00; 0200,:01 cpuinfo:01 userland:01 0200,:01 ocaml:01 uname:01 elephant:98 pae:98 wrote:01 wrote:01 compilers:01 exception:01 caml-list:01 caml-list:01 caml:02 On Tue, Aug 14, 2007 at 06:28:38PM +0200, Koprowski, A. wrote: > Thank you so much for your detailed instructions and help! And > indeed you are completely right, it's not a x86, I don't know what > led me to believe it was. Oh, now you got me all curious. What architecture is it then? Sparc, maybe? I had assumed it was an amd64/x86-64 and that the GNU/Linux installed on it was a 32 bit one. Make me drool a bit, show me /proc/cpuinfo . Ah yes, it makes sense now. Contrary to amd64, as far as I know most GNU/Linux sparc64 installations have a 32 bit userland, but 64 bit kernel, libraries and compilers ready for programs that benefit from it. > -----Original Message----- > From: Richard Jones [mailto:rich@annexia.org] > Sent: Tue 8/14/2007 18:25 > To: Lionel Elie Mamane; Koprowski, A.; caml-list@yquem.inria.fr > Subject: Re: [Caml-list] Utilizing more than 4GB of memory in caml? > > On Tue, Aug 14, 2007 at 06:17:41PM +0200, Lionel Elie Mamane wrote: > > On Tue, Aug 14, 2007 at 04:36:27PM +0200, Koprowski, A. wrote: > > > > >> -----Original Message----- > > >> From: Lionel Elie Mamane [mailto:lionel@mamane.lu] > > > > >> On Tue, Aug 14, 2007 at 04:16:14PM +0200, Koprowski, A. wrote: > > > > >>> I have a 32-bit linux machine with 48-bit addressing of its 128GB > > >>> of RAM. I'd very much like to use this amount of memory; however > > >>> I get Out_of_memory exception after ocaml consumes 4GB. Is there > > >>> anything can do to get my hands on more than > > > > >> Run a 64 bit GNU/Linux on that machine, (...) > > > > > Thanks for the suggestion but I'm afraid I cannot do that. This is > > > a faculty server to which I only have a user access. > > > > Type: > > uname -m > > in a shell. If the answer is x86_64, there are things you can do. If > > it says i386, i486 or i686, all you can do is complain to the system > > administrator (if you are in informatica, I presume that would be bcf > > in room HG 8.73 - 8th floor of Hoofdgebouw; is the machine by any > > chance elephant?). > > > > In case of x86_64: How much manual hacking are you willing to do? > > Is this machine really x86-based? 32-bit x86 machines have at most > 36-bit addresses (through PAE), although that is only usable through > page tables, not to ordinary user processes. 64-bit x86-64 machines > have 48-bit addressing in current incarnations so if they are running > a 32-bit kernel or a 32-bit Xen domain they may fit the description, > but the original poster is still s.o.l. My bet though is it's not x86 > at all. > > Rich. >