From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from sc8-sf-mx1-b.sourceforge.net ([10.3.1.11] helo=sc8-sf-mx1.sourceforge.net) by sc8-sf-list1.sourceforge.net with esmtp (Exim 4.30) id 1BgJdd-0008I0-IS for user-mode-linux-devel@lists.sourceforge.net; Fri, 02 Jul 2004 01:43:17 -0700 Received: from letters.cs.ucsb.edu ([128.111.41.13]) by sc8-sf-mx1.sourceforge.net with esmtp (Exim 4.34) id 1BgJdd-0004nj-9K for user-mode-linux-devel@lists.sourceforge.net; Fri, 02 Jul 2004 01:43:17 -0700 Received: from cs.ucsb.edu (SY197111.resnet.ucsb.edu [128.111.197.111]) by letters.cs.ucsb.edu (8.11.7+Sun/8.11.7) with ESMTP id i628hGT02742 for ; Fri, 2 Jul 2004 01:43:16 -0700 (PDT) Message-ID: <40E51FF0.3020801@cs.ucsb.edu> From: "Lamia M.Youseff" MIME-Version: 1.0 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Subject: [uml-devel] does O_DIRECT hinder greatly file reading performance but not writing?? Sender: user-mode-linux-devel-admin@lists.sourceforge.net Errors-To: user-mode-linux-devel-admin@lists.sourceforge.net List-Unsubscribe: , List-Id: The user-mode Linux development list List-Post: List-Help: List-Subscribe: , List-Archive: Date: Fri, 02 Jul 2004 01:42:24 -0700 To: user-mode-linux-devel@lists.sourceforge.net Hi All,

I am kind of new to UML and I am experiencing a very strange performance hinder with it.
I am running two benchmarks; one that writes a 50Mbyte file and the other one reads a 50Mbyte file.
I tried the benchmarks with and without setting the O_DIRECT flag, in which cases the whole task is done in a matter of seconds(10-30 seconds) except when reading the file with O_DIRECT flag on. In this case, it takes 18-20 min!!!

I am running UML 2.4.19 on top of kernel 2.4.25.  I have time output like that (average of 20 runs)

Operation	User Time (sec)	Sys Time(sec)	Wall Clock Time(sec)
Write Without O_DIRECT	5.612000	0.117500	27.164500
Write With O_DIRECT	10.007000	0.017000	22.199500
Read without O_DIRECT	3.035000	0.091000	7.760000
Read With O_DIRECT	21.492000	29.680000	1133.054000

I would appreciate it very much if somebody would explain to me how to solve that strange performance problem, or even explain why does it happen for reading with O_DIRECT but not without !!??

Thanks allot,
Lamia Youseff
--

I am kind of new to UML and I am experiencing a very strange performance hinder with it.
I am running two benchmarks; one that writes a 50Mbyte file and the other one reads a 50Mbyte file.
I tried the benchmarks with and without setting the O_DIRECT flag, in which cases the whole task is done in a matter of seconds(10-30 seconds) except when reading the file with O_DIRECT flag on. In this case, it takes 18-20 min!!!

I am running UML 2.4.19 on top of kernel 2.4.25.  I have time output like that (average of 20 runs)

Operation	User Time (sec)	Sys Time(sec)	Wall Clock Time(sec)
Write Without O_DIRECT	5.612000	0.117500	27.164500
Write With O_DIRECT	10.007000	0.017000	22.199500
Read without O_DIRECT	3.035000	0.091000	7.760000
Read With O_DIRECT	21.492000	29.680000	1133.054000

I would appreciate it very much if somebody would explain to me how to solve that strange performance problem, or even explain why does it happen for reading with O_DIRECT but not without !!??

Thanks allot,
Lamia Youseff
--

------------------------------------------------------- This SF.Net email sponsored by Black Hat Briefings & Training. Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com _______________________________________________ User-mode-linux-devel mailing list User-mode-linux-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from sc8-sf-mx2-b.sourceforge.net ([10.3.1.12] helo=sc8-sf-mx2.sourceforge.net) by sc8-sf-list1.sourceforge.net with esmtp (Exim 4.30) id 1Bh3ni-0002Wn-O9 for user-mode-linux-devel@lists.sourceforge.net; Sun, 04 Jul 2004 03:00:46 -0700 Received: from smtp002.mail.ukl.yahoo.com ([217.12.11.33]) by sc8-sf-mx2.sourceforge.net with smtp (Exim 4.34) id 1Bh3ni-0001R1-6N for user-mode-linux-devel@lists.sourceforge.net; Sun, 04 Jul 2004 03:00:46 -0700 From: BlaisorBlade Subject: Please upgrade! (was: Re: [uml-devel] does O_DIRECT hinder greatly file reading performance but not writing??) References: <40E51FF0.3020801@cs.ucsb.edu> In-Reply-To: <40E51FF0.3020801@cs.ucsb.edu> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Message-Id: <200407032128.40746.blaisorblade_spam@yahoo.it> Sender: user-mode-linux-devel-admin@lists.sourceforge.net Errors-To: user-mode-linux-devel-admin@lists.sourceforge.net List-Unsubscribe: , List-Id: The user-mode Linux development list List-Post: List-Help: List-Subscribe: , List-Archive: Date: Sat, 3 Jul 2004 21:28:40 +0200 Content-Transfer-Encoding: quoted-printable To: "Lamia M.Youseff" , user-mode-linux-devel@lists.sourceforge.net Alle 10:42, venerd=EC 2 luglio 2004, Lamia M.Youseff ha scritto: > I am running UML 2.4.19 on top of kernel 2.4.25. I have time output like > that (average of 20 runs)=20 Well, could you try upgrading it to 2.4.24-1um and retesting? That UML is o= ld,=20 buggy, and outdated. --=20 Paolo Giarrusso, aka Blaisorblade Linux registered user n. 292729 ------------------------------------------------------- This SF.Net email sponsored by Black Hat Briefings & Training. Attend Black Hat Briefings & Training, Las Vegas July 24-29 -=20 digital self defense, top technical experts, no vendor pitches,=20 unmatched networking opportunities. Visit www.blackhat.com _______________________________________________ User-mode-linux-devel mailing list User-mode-linux-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from sc8-sf-mx2-b.sourceforge.net ([10.3.1.12] helo=sc8-sf-mx2.sourceforge.net) by sc8-sf-list1.sourceforge.net with esmtp (Exim 4.30) id 1Bhib5-00065z-1I for user-mode-linux-devel@lists.sourceforge.net; Mon, 05 Jul 2004 22:34:27 -0700 Received: from letters.cs.ucsb.edu ([128.111.41.13]) by sc8-sf-mx2.sourceforge.net with esmtp (Exim 4.34) id 1Bhib4-0004Nt-Mg for user-mode-linux-devel@lists.sourceforge.net; Mon, 05 Jul 2004 22:34:26 -0700 Message-ID: <40EA39DC.5030409@cs.ucsb.edu> From: "Lamia M.Youseff" MIME-Version: 1.0 Subject: Re: Please upgrade! (was: Re: [uml-devel] does O_DIRECT hinder greatly file reading performance but not writing??) References: <40E51FF0.3020801@cs.ucsb.edu> <200407032128.40746.blaisorblade_spam@yahoo.it> Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: user-mode-linux-devel-admin@lists.sourceforge.net Errors-To: user-mode-linux-devel-admin@lists.sourceforge.net List-Unsubscribe: , List-Id: The user-mode Linux development list List-Post: List-Help: List-Subscribe: , List-Archive: Date: Mon, 05 Jul 2004 22:34:20 -0700 To: BlaisorBlade Cc: user-mode-linux-devel@lists.sourceforge.net Hello All, Hi BlaisorBlade,
I upgraded to 2.4.24-1um, but the same error still persist on the um..i.e. the disk reads takes so long when used with the O_DIRECT flag !!
Anybody has any explanation for that??

Thank you,
Lamia Youseff

BlaisorBlade wrote:

Alle 10:42, venerdì 2 luglio 2004, Lamia M.Youseff ha scritto:

I am running UML 2.4.19 on top of kernel 2.4.25.  I have time output like
that (average of 20 runs)

Well, could you try upgrading it to 2.4.24-1um and retesting? That UML is old, 
buggy, and outdated.

-- 
..........................................................
: Lamia M.Youseff          :  lyouseff@cs.ucsb.edu       :
: Ph.D student             :  www.cs.ucsb.edu/~lyouseff  :
: University of California,:                             :
: Santa Barbara (UCSB)     :                             :
:........................................................:
:                                                        :
: "There are no impossible dreams; there is just our     :
: limited perception of what is possible"                :
*:-.,_,.-:*'``'*:-.,_,.-:*'``'*:-.,_,.-:*'``'*:-.,_,.-:*'*