From wm_freesky at 126.com  Mon Nov  5 09:00:40 2007
From: wm_freesky at 126.com (Wang Min)
Date: Mon, 5 Nov 2007 22:00:40 +0800
Subject: [Disksim-users] Why is SysSim time negative value?
Message-ID: <006701c81fb4$41c12350$6b280a0a@ictwangmin>

Hi,all:
     I'm going to run the system-level simulation with syssim_driver.c, but 
the response time is a negative value -39163.548316,   why? I have also read 
through the related code, but can't explain it.

    syssim: externally-controlled DiskSim (avg. resp should be about 19.6ms)
  response time: n=1000 average=-39163.548316 std. deviation=22560.393800
IOdriver Response time average:         19.561833

PS:  the command is : ../src/syssim  st41601n.parv  syssim.outv  2676846  0 





From wm_freesky at 126.com  Mon Nov  5 20:05:58 2007
From: wm_freesky at 126.com (Wang Min)
Date: Tue, 6 Nov 2007 09:05:58 +0800
Subject: [Disksim-users] Why is SysSim time negative value?
References: <006701c81fb4$41c12350$6b280a0a@ictwangmin>
Message-ID: <002901c82011$30f2cad0$6b280a0a@ictwangmin>

hi, all:
     Maybe I didn't describe the question clearly. I run the runvalid script 
in DiskSim, and the last two validations are  about system-level simulation 
with DiskSim. I am confused on this results: IOdriver response time is 
19.56ms, but the total system response time is -39163.54.  Would you please 
explain it to me? Thank you very much.

The results is :
syssim: externally-controlled DiskSim (avg. resp should be about 19.6ms)
  response time: n=1000 average=-39163.548316 std. deviation=22560.393800
IOdriver Response time average:         19.561833

yours  MinWang



----- Original Message ----- 
From: "Wang Min" <wm_freesky at 126.com>
To: <disksim-users at ece.cmu.edu>
Sent: Monday, November 05, 2007 10:00 PM
Subject: [Disksim-users] Why is SysSim time negative value?


> Hi,all:
>     I'm going to run the system-level simulation with syssim_driver.c, but 
> the response time is a negative value -39163.548316,   why? I have also 
> read through the related code, but can't explain it.
>
>    syssim: externally-controlled DiskSim (avg. resp should be about 
> 19.6ms)
>  response time: n=1000 average=-39163.548316 std. deviation=22560.393800
> IOdriver Response time average:         19.561833
>
> PS:  the command is : ../src/syssim  st41601n.parv  syssim.outv  2676846 
> 0
>
>
> _______________________________________________
> Disksim-users mailing list
> Disksim-users at ece.cmu.edu
> https://sos.ece.cmu.edu/mailman/listinfo/disksim-users
> 





From keepeffort at naver.com  Wed Nov  7 10:23:26 2007
From: keepeffort at naver.com (=?US-ASCII?B?a2VlcGVmZm9ydA==?=)
Date: Thu, 08 Nov 2007 00:23:26 +0900
Subject: [Disksim-users] param file
Message-ID: <a3798109be62defaa686f871c85f00b8@i61797>

Hi all,
 
I have a question about param file.
I made a param file by 'vi editor' and I simulated a simple trace file.
But it just error message is printed that is [Error reading 'Byte Order:'].
Apparently, I inputted 'Little' at param file, nevertheless, for what reason printed this error message.
Is that wrong what form of my param file made? If so, how to make a param file?
Thanks alot!
 
Best,
YG
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.andrew.cmu.edu/pipermail/disksim-users/attachments/20071108/02984a00/attachment.html>

From keepeffort at naver.com  Tue Nov 13 12:42:00 2007
From: keepeffort at naver.com (=?US-ASCII?B?a2VlcGVmZm9ydA==?=)
Date: Wed, 14 Nov 2007 02:42:00 +0900
Subject: [Disksim-users] Error reading 'Byte Order:'
Message-ID: <8f9c6852edfb01c220e18d6ed0a5f1af@i61825>

Hi all,
 
Line 384 to 395 in Disksim.c, 
 
384 if (fscanf(parfile, "\nByte Order (Endian): %s\n", seed) != 1) {
385 fprintf(stderr, "Error reading 'Byte Order:'\n");
386 exit(0);
387 }
388 if (strcmp(seed, "Little") == 0) {
389 disksim->endian = _LITTLE_ENDIAN;
390 } else if (strcmp(seed, "Big") == 0) {
391 disksim->endian = _BIG_ENDIAN;
392 } else {
393 fprintf(stderr, "Invalid byte ordering scheme - %s\n", seed);
394 exit(0);
395 }
 
'Fscanf' scan a stored data at parfile. And then, it compare the string whether 'Little' or 'Big'.
Now, 'Little' is stored at 'parfile', but it happen error at line number of 385 . Error reading 'Byte Order:'.
I don't know why this error is happened.
I'll very appreciate of any advice.
 
Best regards,
YG
 
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.andrew.cmu.edu/pipermail/disksim-users/attachments/20071114/43bd8af0/attachment.html>

From chemical.aniket at gmail.com  Fri Nov 16 02:20:35 2007
From: chemical.aniket at gmail.com (aniket)
Date: Fri, 16 Nov 2007 01:20:35 -0600
Subject: [Disksim-users] aniket is your Yaar! :)
Message-ID: <a4fcb90252c8911287d5b11decca0f9c@localhost.localdomain>

An HTML attachment was scrubbed...
URL: <http://lists.andrew.cmu.edu/pipermail/disksim-users/attachments/20071116/a1bab329/attachment.html>

From disksim-users at ece.cmu.edu  Mon Nov 19 11:08:29 2007
From: disksim-users at ece.cmu.edu (VIAGRA ® Official Site)
Date: Mon, 19 Nov 2007 11:08:29 -0500 (EST)
Subject: [Disksim-users] November 73% OFF
Message-ID: <20071119060829.3220.qmail@dslb-088-068-113-029.pools.arcor-ip.net>

An HTML attachment was scrubbed...
URL: <http://lists.andrew.cmu.edu/pipermail/disksim-users/attachments/20071119/b555f991/attachment.html>

From smita.vishwa at gmail.com  Wed Nov 21 00:34:45 2007
From: smita.vishwa at gmail.com (smita vishwakarma)
Date: Wed, 21 Nov 2007 11:04:45 +0530
Subject: [Disksim-users] Use of Disk simulator with NS2
Message-ID: <df1e80350711202134o17f9c60co971d2a9dceecf757@mail.gmail.com>

Dear all,

I am going to start the simulation of disk with the Network Simulator
but, I don't know how to use the Disk Simulator with NS2 . Please ,
help me on this.

Thanks & Regards,
Smita Vishwakarma


From maheshwari at wisc.edu  Sat Nov 24 22:31:54 2007
From: maheshwari at wisc.edu (MAYANK MAHESHWARI)
Date: Sat, 24 Nov 2007 21:31:54 -0600
Subject: [Disksim-users] Controller Interface
Message-ID: <ea27c8da7cf4.4748984a@wiscmail.wisc.edu>

Hi,

I'm trying to locate the modules in the source code which schedules the
requests and passes the writes to the bus or the driver. Basic logical
hierarchy suggests that its the controller-driver-bus-disk . So if I want
to give more than one addresses to write to, will disksim be able to
handle that (there will have to be changes made) and which modules would
likely be needed to be hacked?

Mayank.



From ganger at ece.cmu.edu  Sat Nov 24 22:50:08 2007
From: ganger at ece.cmu.edu (Greg Ganger)
Date: Sat, 24 Nov 2007 22:50:08 -0500 (EST)
Subject: [Disksim-users] Controller Interface
In-Reply-To: <ea27c8da7cf4.4748984a@wiscmail.wisc.edu>
References: <ea27c8da7cf4.4748984a@wiscmail.wisc.edu>
Message-ID: <Pine.LNX.4.64.0711242248370.22677@nirvana.pdl.cmu.edu>


I don't understand the question, but I believe that the user manual
describes the hierarchy of components.

Greg


On Sat, 24 Nov 2007, MAYANK MAHESHWARI wrote:

> Hi,
>
> I'm trying to locate the modules in the source code which schedules the
> requests and passes the writes to the bus or the driver. Basic logical
> hierarchy suggests that its the controller-driver-bus-disk . So if I want
> to give more than one addresses to write to, will disksim be able to
> handle that (there will have to be changes made) and which modules would
> likely be needed to be hacked?
>
> Mayank.
>
> _______________________________________________
> Disksim-users mailing list
> Disksim-users at ece.cmu.edu
> https://sos.ece.cmu.edu/mailman/listinfo/disksim-users
>


From maheshwari at wisc.edu  Sat Nov 24 23:11:18 2007
From: maheshwari at wisc.edu (MAYANK MAHESHWARI)
Date: Sat, 24 Nov 2007 22:11:18 -0600
Subject: [Disksim-users] Controller Interface
Message-ID: <e76af208637d.4748a186@wiscmail.wisc.edu>

I understand the hierarchy and its given in the manual. What i was interested in was if I want to provide Disksim a certain range of addresses to write to instead of a single block number, where would the changes be required to make: in the controller or the IO driver. My guess is it should be somewhere in the controller...so I would wanna know where does the scheduling take place in the controller and if by making changes to the scheduler, could you achieve writes which are addressed to a range of physical addresses? 


From ganger at ece.cmu.edu  Sat Nov 24 23:53:20 2007
From: ganger at ece.cmu.edu (Greg Ganger)
Date: Sat, 24 Nov 2007 23:53:20 -0500 (EST)
Subject: [Disksim-users] Controller Interface
In-Reply-To: <e76af208637d.4748a186@wiscmail.wisc.edu>
References: <e76af208637d.4748a186@wiscmail.wisc.edu>
Message-ID: <Pine.LNX.4.64.0711242351080.22677@nirvana.pdl.cmu.edu>


Depends on where you wanted it to happen.  Think of it like an actual
system, which could do that in any of several (or multiple) places.
For example, scheduling happens in many places -- scheduling parameters
can be provided for driver, controller, and disk (see manual).

Good luck,

Greg


On Sat, 24 Nov 2007, MAYANK MAHESHWARI wrote:

> I understand the hierarchy and its given in the manual. What i was interested in was if I want to provide Disksim a certain range of addresses to write to instead of a single block number, where would the changes be required to make: in the controller or the IO driver. My guess is it should be somewhere in the controller...so I would wanna know where does the scheduling take place in the controller and if by making changes to the scheduler, could you achieve writes which are addressed to a range of physical addresses?
> _______________________________________________
> Disksim-users mailing list
> Disksim-users at ece.cmu.edu
> https://sos.ece.cmu.edu/mailman/listinfo/disksim-users
>


From markus at bluegap.ch  Tue Nov 27 07:00:08 2007
From: markus at bluegap.ch (Markus Schiltknecht)
Date: Tue, 27 Nov 2007 13:00:08 +0100
Subject: [Disksim-users] status of disksim
Message-ID: <474C06C8.5070505@bluegap.ch>

Hi,

I'm wondering about the status of disksim. I've read quite some requests 
for data for newer drives - can somebody provide some? Is the CVS of 
disksim publicly accessible? Is there a new release planned?

Regards

Markus




From markus at bluegap.ch  Tue Nov 27 14:51:17 2007
From: markus at bluegap.ch (Markus Schiltknecht)
Date: Tue, 27 Nov 2007 20:51:17 +0100
Subject: [Disksim-users] status of disksim
In-Reply-To: <474C06C8.5070505@bluegap.ch>
References: <474C06C8.5070505@bluegap.ch>
Message-ID: <474C7535.7000301@bluegap.ch>

Hi,

Markus Schiltknecht wrote:
> I'm wondering about the status of disksim. I've read quite some requests 
> for data for newer drives - can somebody provide some? Is the CVS of 
> disksim publicly accessible? Is there a new release planned?

... is Dixtract available somewhere?

I've just managed to compile the source under Linux. And I've some 
compiler warnings. I'd like to make disksim simpler to compile by 
introducing some sort of autoconf magic.

Regards

Markus



From michael.meeks at novell.com  Wed Nov 28 03:36:45 2007
From: michael.meeks at novell.com (Michael Meeks)
Date: Wed, 28 Nov 2007 08:36:45 +0000
Subject: [Disksim-users] status of disksim
In-Reply-To: <474C7535.7000301@bluegap.ch>
References: <474C06C8.5070505@bluegap.ch>  <474C7535.7000301@bluegap.ch>
Message-ID: <1196239005.31118.28.camel@localhost.localdomain>

Hi Markus,

On Tue, 2007-11-27 at 20:51 +0100, Markus Schiltknecht wrote:
> I've just managed to compile the source under Linux. And I've some 
> compiler warnings. I'd like to make disksim simpler to compile by 
> introducing some sort of autoconf magic.

	This has already been done as part of iogrind, with a load of bug
fixes, 64bit fixes etc.

	See:

	http://live.gnome.org/iogrind/

	for the source download URL; I'd be happy to see someone maintaining
that ;-)

	HTH,

		Michael.

-- 
 michael.meeks at novell.com  <><, Pseudo Engineer, itinerant idiot




From markus at bluegap.ch  Wed Nov 28 09:07:30 2007
From: markus at bluegap.ch (Markus Schiltknecht)
Date: Wed, 28 Nov 2007 15:07:30 +0100
Subject: [Disksim-users] status of disksim
In-Reply-To: <1196239005.31118.28.camel@localhost.localdomain>
References: <474C06C8.5070505@bluegap.ch> <474C7535.7000301@bluegap.ch>
	<1196239005.31118.28.camel@localhost.localdomain>
Message-ID: <474D7622.1090609@bluegap.ch>

Hello Michael,

Michael Meeks wrote:
> 	This has already been done as part of iogrind, with a load of bug
> fixes, 64bit fixes etc.
> 
> 	See:
> 
> 	http://live.gnome.org/iogrind/

Thank you for that pointer, interesting project.

You don't mention disksim at all on your iogrind website. Why do you 
clam it's inaccurate, while disksim is said to be quite accurate? Did 
you drop certain capabilities of disksim or is this because of your 
broken 100Gb disk model?

Regards

Markus


From bucy at gloop.org  Wed Nov 28 16:41:37 2007
From: bucy at gloop.org (John Bucy)
Date: Wed, 28 Nov 2007 13:41:37 -0800
Subject: [Disksim-users] status of disksim
In-Reply-To: <474D7622.1090609@bluegap.ch>
References: <474C06C8.5070505@bluegap.ch> <474C7535.7000301@bluegap.ch>
	<1196239005.31118.28.camel@localhost.localdomain>
	<474D7622.1090609@bluegap.ch>
Message-ID: <4fc1e0430711281341gb9d381bia5694b74fb421fcb@mail.gmail.com>

Hi all,

Our group of CMU/PDL alumni are still trying to get a new disksim
release together;  we're just very busy with our day jobs and it's
been hard to make time to work on it.  I believe we're down to the
last one or two outstanding issues that need to be fixed before we can
ship so we may actually be able to get it done this month!



john

On Nov 28, 2007 6:07 AM, Markus Schiltknecht <markus at bluegap.ch> wrote:
> Hello Michael,
>
> Michael Meeks wrote:
> >       This has already been done as part of iogrind, with a load of bug
> > fixes, 64bit fixes etc.
> >
> >       See:
> >
> >       http://live.gnome.org/iogrind/
>
> Thank you for that pointer, interesting project.
>
> You don't mention disksim at all on your iogrind website. Why do you
> clam it's inaccurate, while disksim is said to be quite accurate? Did
> you drop certain capabilities of disksim or is this because of your
> broken 100Gb disk model?
>
> Regards
>
> Markus
>
> _______________________________________________
> Disksim-users mailing list
> Disksim-users at ece.cmu.edu
> https://sos.ece.cmu.edu/mailman/listinfo/disksim-users
>


From michael.meeks at novell.com  Thu Nov 29 01:19:43 2007
From: michael.meeks at novell.com (Michael Meeks)
Date: Thu, 29 Nov 2007 06:19:43 +0000
Subject: [Disksim-users] status of disksim
In-Reply-To: <474D7622.1090609@bluegap.ch>
References: <474C06C8.5070505@bluegap.ch>  <474C7535.7000301@bluegap.ch>
	<1196239005.31118.28.camel@localhost.localdomain>
	<474D7622.1090609@bluegap.ch>
Message-ID: <1196317183.17150.18.camel@localhost.localdomain>


On Wed, 2007-11-28 at 15:07 +0100, Markus Schiltknecht wrote:
> You don't mention disksim at all on your iogrind website.

	That is a bug, please do fix the wiki page to add a link (I could have
sworn it was there, it's prominently mentioned in the source README
etc.).

>  Why do you clam it's inaccurate, while disksim is said to be
> quite accurate? Did you drop certain capabilities of disksim or
> is this because of your broken 100Gb disk model?

	My 100Gb disk model, produced by tweaking the umpteen tweakables in
there. In the process of tweaking them, some pretty anomylous results
were generated, leading me to suspect some rather un-tested and
under-constrained validation of that file. Of course, my 100Gb disk is
all one zone (eg.) ;-)

	Of course, using the more simple disk models - such as constant seek
time data was not useful for what I wanted ... anyhow ;-) if someone
that really understands all the tweakables can produce a reasonable
model of a reasonably modern disk that'd be most helpful.

	Regards,

		Michael.

-- 
 michael.meeks at novell.com  <><, Pseudo Engineer, itinerant idiot




From stroucki at ece.cmu.edu  Fri Nov 30 11:16:08 2007
From: stroucki at ece.cmu.edu (Michael Stroucken)
Date: Fri, 30 Nov 2007 11:16:08 -0500
Subject: [Disksim-users] Why DiskSIM not support the SRT V1.6 formated trace
	file?
Message-ID: <47503748.7040807@ece.cmu.edu>

Original message from kejian at nrchpc.ac.cn. Forwarding message to 
disksim-users -- Michael.

Hi?all

  When I use the cello99 trace file from HPL in DiskSIM, the error 
message is:



?Format problem with 'tracedate' line in HPL trace - tracedate = Mon 
Feb  1 12:00:00 1999?



The cello92 trace file is OK. Who can tell me how to modify the DiskSIM?