Quantcast

RE: Linux - Odd results when compared to Windows 2 003

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Linux - Odd results when compared to Windows 2 003

Matthew Wilby
RE: [Iometer-user] Linux - Odd results when compared to Windows 2003

* what is you HW spec?

Test machine runs dual P4 3.2 Ghz Xeons. 2GB RAM. Two identical SCSI HDDs, one for Windows 2003 and the other for Linux.

* when test a U320 drive, you test raw disk or a disk with file system?

Disk with filesystem. Partitioned to 250MB & formatted with Ext3. Same size partition on Windows but obviously formatted with NTFS.

* if test on file system, what is the size of your testing file?

We just used the entire 250MB partition both in Windows & Linux.

* what iometer version u are using?

iometer-2004.07.30


-----Original Message-----
From: Ming Zhang
To: Matthew Wilby
Cc: iometer-user-list
Sent: 27/09/2005 18:55
Subject: Re: [Iometer-user] Linux - Odd results when compared to Windows 2003

On Tue, 2005-09-27 at 17:27 +0100, Matthew Wilby wrote:
> I've been running some basic disk benchmarking for local SCSI U320, FC
> & iSCSI attached storage, to compare performance under Windows &
> Linux.
>
> All tests have been carried out using the same equipment, and same
> test setup. The only difference is the OS. The test bed server is dual
> boot Win2k3 and Linux (SuSE SLES 9 - 2.6.5 kernel). 
>

* what is you HW spec?
* when test a U320 drive, you test raw disk or a disk with file system?
* if test on file system, what is the size of your testing file?
* what iometer version u are using?



> The results I get back under Linux are bizarre! Below are a few of the
> results......


> 100% rand. write
>
> Windows 2003 - 27 439
> Linux - 629 10071
>
>
> 100% seq. write
>
> Windows 2003 - 67 1068
> Linux - 627 100033
>
>
> 100% rand. read
>
> Windows 2003 - 30 491
> Linux - 1777 28431
>
>
> 100% seq. read
>
> Windows 2003 - 67 1079
> Linux - 1769 28312
>
>
> 20% read 80% write
> (50% seq./write)
>
> Windows 2003 - 32 517
> Linux - 687 10996
>
>
> 50% read/write
> (50% seq./write)
>
> Windows 2003 - 26 419
> Linux - 903 14452
>

> The first figure for each test is MB/sec and the second in IOPS. These
> are tests from local SCSI U320 drives.

> How can I accurately benchmark I/O between Windows & Linux with
> IOmeter? Is this a bug or can these figures be explained?

> Thanks.

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Linux - Odd results when compared to Windows 2 003

Ming Zhang
On Tue, 2005-09-27 at 21:41 +0100, Matthew Wilby wrote:

> * what is you HW spec?
>
> Test machine runs dual P4 3.2 Ghz Xeons. 2GB RAM. Two identical SCSI
> HDDs, one for Windows 2003 and the other for Linux.
>
> * when test a U320 drive, you test raw disk or a disk with file
> system?
>
> Disk with filesystem. Partitioned to 250MB & formatted with Ext3. Same
> size partition on Windows but obviously formatted with NTFS.

this is the problem. u have 250MB partition while 2G ram. so most w/r
will happen in RAM totally. at least use 5X u ram size space.


>
> * if test on file system, what is the size of your testing file?
>
> We just used the entire 250MB partition both in Windows & Linux.
>
> * what iometer version u are using?
>
> iometer-2004.07.30
>
>
> -----Original Message-----
> From: Ming Zhang
> To: Matthew Wilby
> Cc: iometer-user-list
> Sent: 27/09/2005 18:55
> Subject: Re: [Iometer-user] Linux - Odd results when compared to
> Windows 2003
>
> On Tue, 2005-09-27 at 17:27 +0100, Matthew Wilby wrote:
> > I've been running some basic disk benchmarking for local SCSI U320,
> FC
> > & iSCSI attached storage, to compare performance under Windows &
> > Linux.  
> >  
> > All tests have been carried out using the same equipment, and same
> > test setup. The only difference is the OS. The test bed server is
> dual
> > boot Win2k3 and Linux (SuSE SLES 9 - 2.6.5 kernel).  
> >
>
> * what is you HW spec?
> * when test a U320 drive, you test raw disk or a disk with file
> system?
> * if test on file system, what is the size of your testing file?
> * what iometer version u are using?
>
>
>
> > The results I get back under Linux are bizarre! Below are a few of
> the
> > results......
> >  
> >  
> > 100% rand. write  
> >  
> > Windows 2003 - 27 439  
> > Linux - 629 10071  
> >  
> >  
> > 100% seq. write  
> >  
> > Windows 2003 - 67 1068  
> > Linux - 627 100033  
> >  
> >  
> > 100% rand. read  
> >  
> > Windows 2003 - 30 491  
> > Linux - 1777 28431  
> >  
> >  
> > 100% seq. read  
> >  
> > Windows 2003 - 67 1079  
> > Linux - 1769 28312  
> >  
> >  
> > 20% read 80% write  
> > (50% seq./write)  
> >  
> > Windows 2003 - 32 517  
> > Linux - 687 10996  
> >  
> >  
> > 50% read/write  
> > (50% seq./write)  
> >  
> > Windows 2003 - 26 419  
> > Linux - 903 14452  
> >  
> >  
> > The first figure for each test is MB/sec and the second in IOPS.
> These
> > are tests from local SCSI U320 drives.
> >  
> > How can I accurately benchmark I/O between Windows & Linux with
> > IOmeter? Is this a bug or can these figures be explained?
> >  
> > Thanks.
>



-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
_______________________________________________
Iometer-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/iometer-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Linux - Odd results when compared to Windows 2 003

Matthew Wilby
In reply to this post by Matthew Wilby
RE: [Iometer-user] Linux - Odd results when compared to Windows 2003

Kate,

We have 2GB RAM in the test box + 2GB swap partiton in Linux/2GB page file in Windows. Note: this is on the same box.

As you can see from the results, the figures in Windows are sensible, whereas the figures in Linux are not.

How would I disable caching within Linux? 

Thanks,

Matt


-----Original Message-----
From: Coyne, Kate
To: Matthew Wilby; [hidden email]
Sent: 27/09/2005 18:09
Subject: RE: [Iometer-user] Linux - Odd results when compared to Windows 2003

Matthew,

 I looks like, on Linux, you have a good amount of cache on the system
which would effect the iometer on linux numbers to be inflated.  Did you
turn off the cache(system) so that it's more a raw disk test.  I was
having problems with Random Writes on Linux and have not been able to
get good consistant numbers here...and I only have 512MB of RAM so that
there is minimal cache hits plus I use a 2Gb file.

Hope this helps.

Kate

 

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Linux - Odd results when compared to Windows 2 003

Matthew Wilby
In reply to this post by Matthew Wilby
RE: [Iometer-user] Linux - Odd results when compared to Windows 2 003

"this is the problem. u have 250MB partition while 2G ram. so most w/r
will happen in RAM totally. at least use 5X u ram size space. "

Is this rule also true for Windows, or just Linux/Unix? If so, I'm assuming Windows & Linux behave differently when swriting to disk?

Thanks,

Matt


-----Original Message-----
From: Ming Zhang
To: Matthew Wilby
Cc: 'iometer-user-list '
Sent: 27/09/2005 21:48
Subject: RE: [Iometer-user] Linux - Odd results when compared to Windows 2 003

On Tue, 2005-09-27 at 21:41 +0100, Matthew Wilby wrote:
> * what is you HW spec?
>
> Test machine runs dual P4 3.2 Ghz Xeons. 2GB RAM. Two identical SCSI
> HDDs, one for Windows 2003 and the other for Linux.
>
> * when test a U320 drive, you test raw disk or a disk with file
> system?
>
> Disk with filesystem. Partitioned to 250MB & formatted with Ext3. Same
> size partition on Windows but obviously formatted with NTFS.

this is the problem. u have 250MB partition while 2G ram. so most w/r
will happen in RAM totally. at least use 5X u ram size space.


>
> * if test on file system, what is the size of your testing file?
>
> We just used the entire 250MB partition both in Windows & Linux.
>
> * what iometer version u are using?
>
> iometer-2004.07.30
>
>
> -----Original Message-----
> From: Ming Zhang
> To: Matthew Wilby
> Cc: iometer-user-list
> Sent: 27/09/2005 18:55
> Subject: Re: [Iometer-user] Linux - Odd results when compared to
> Windows 2003
>
> On Tue, 2005-09-27 at 17:27 +0100, Matthew Wilby wrote:
> > I've been running some basic disk benchmarking for local SCSI U320,
> FC
> > & iSCSI attached storage, to compare performance under Windows &
> > Linux. 
> > 
> > All tests have been carried out using the same equipment, and same
> > test setup. The only difference is the OS. The test bed server is
> dual
> > boot Win2k3 and Linux (SuSE SLES 9 - 2.6.5 kernel).  
> >
>
> * what is you HW spec?
> * when test a U320 drive, you test raw disk or a disk with file
> system?
> * if test on file system, what is the size of your testing file?
> * what iometer version u are using?
>
>
>
> > The results I get back under Linux are bizarre! Below are a few of
> the
> > results......
> >  
> >  
> > 100% rand. write 
> > 
> > Windows 2003 - 27 439 
> > Linux - 629 10071 
> > 
> > 
> > 100% seq. write 
> > 
> > Windows 2003 - 67 1068 
> > Linux - 627 100033 
> > 
> > 
> > 100% rand. read 
> > 
> > Windows 2003 - 30 491 
> > Linux - 1777 28431 
> > 
> > 
> > 100% seq. read 
> > 
> > Windows 2003 - 67 1079 
> > Linux - 1769 28312 
> > 
> > 
> > 20% read 80% write 
> > (50% seq./write) 
> > 
> > Windows 2003 - 32 517 
> > Linux - 687 10996 
> > 
> > 
> > 50% read/write 
> > (50% seq./write) 
> > 
> > Windows 2003 - 26 419 
> > Linux - 903 14452 
> > 
> >  
> > The first figure for each test is MB/sec and the second in IOPS.
> These
> > are tests from local SCSI U320 drives.
> >  
> > How can I accurately benchmark I/O between Windows & Linux with
> > IOmeter? Is this a bug or can these figures be explained?
> >  
> > Thanks.
>

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Linux - Odd results when compared to Windows 2 003

Ming Zhang
On Tue, 2005-09-27 at 21:58 +0100, Matthew Wilby wrote:
> "this is the problem. u have 250MB partition while 2G ram. so most
> w/r
> will happen in RAM totally. at least use 5X u ram size space. "
>
> Is this rule also true for Windows, or just Linux/Unix? If so, I'm
> assuming Windows & Linux behave differently when swriting to disk?
>

had better for both. but as you saw, your windows test can get sensible
results. i work more on linux, and i know that is one of the problem.



> Thanks,
>
> Matt
>
>
> -----Original Message-----
> From: Ming Zhang
> To: Matthew Wilby
> Cc: 'iometer-user-list '
> Sent: 27/09/2005 21:48
> Subject: RE: [Iometer-user] Linux - Odd results when compared to
> Windows 2 003
>
> On Tue, 2005-09-27 at 21:41 +0100, Matthew Wilby wrote:
> > * what is you HW spec?
> >  
> > Test machine runs dual P4 3.2 Ghz Xeons. 2GB RAM. Two identical
> SCSI
> > HDDs, one for Windows 2003 and the other for Linux.  
> >  
> > * when test a U320 drive, you test raw disk or a disk with file
> > system?
> >  
> > Disk with filesystem. Partitioned to 250MB & formatted with Ext3.
> Same
> > size partition on Windows but obviously formatted with NTFS.
>
> this is the problem. u have 250MB partition while 2G ram. so most w/r
> will happen in RAM totally. at least use 5X u ram size space.
>
>
> >  
> > * if test on file system, what is the size of your testing file?
> >  
> > We just used the entire 250MB partition both in Windows & Linux.
> >  
> > * what iometer version u are using?  
> >  
> > iometer-2004.07.30
> >  
> >  
> > -----Original Message-----  
> > From: Ming Zhang  
> > To: Matthew Wilby  
> > Cc: iometer-user-list  
> > Sent: 27/09/2005 18:55  
> > Subject: Re: [Iometer-user] Linux - Odd results when compared to
> > Windows 2003
> >  
> > On Tue, 2005-09-27 at 17:27 +0100, Matthew Wilby wrote:  
> > > I've been running some basic disk benchmarking for local SCSI
> U320,
> > FC  
> > > & iSCSI attached storage, to compare performance under Windows &  
> > > Linux.  
> > >  
> > > All tests have been carried out using the same equipment, and
> same  
> > > test setup. The only difference is the OS. The test bed server is
> > dual  
> > > boot Win2k3 and Linux (SuSE SLES 9 - 2.6.5 kernel).    
> > >  
> >  
> > * what is you HW spec?  
> > * when test a U320 drive, you test raw disk or a disk with file
> > system?  
> > * if test on file system, what is the size of your testing file?  
> > * what iometer version u are using?
> >  
> >  
> >  
> > > The results I get back under Linux are bizarre! Below are a few
> of
> > the  
> > > results......  
> > >    
> > >    
> > > 100% rand. write  
> > >  
> > > Windows 2003 - 27 439  
> > > Linux - 629 10071  
> > >  
> > >  
> > > 100% seq. write  
> > >  
> > > Windows 2003 - 67 1068  
> > > Linux - 627 100033  
> > >  
> > >  
> > > 100% rand. read  
> > >  
> > > Windows 2003 - 30 491  
> > > Linux - 1777 28431  
> > >  
> > >  
> > > 100% seq. read  
> > >  
> > > Windows 2003 - 67 1079  
> > > Linux - 1769 28312  
> > >  
> > >  
> > > 20% read 80% write  
> > > (50% seq./write)  
> > >  
> > > Windows 2003 - 32 517  
> > > Linux - 687 10996  
> > >  
> > >  
> > > 50% read/write  
> > > (50% seq./write)  
> > >  
> > > Windows 2003 - 26 419  
> > > Linux - 903 14452  
> > >  
> > >    
> > > The first figure for each test is MB/sec and the second in IOPS.
> > These  
> > > are tests from local SCSI U320 drives.  
> > >    
> > > How can I accurately benchmark I/O between Windows & Linux with  
> > > IOmeter? Is this a bug or can these figures be explained?  
> > >    
> > > Thanks.
> >
>



-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
_______________________________________________
Iometer-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/iometer-user
Loading...