Native Command Queuing w/ K8N Neo2

[jr]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

I just enabled Native Command Queuing on my 74GB WD Raptor . I was curious
if anyone with the same setup has seen any improvements (particularly with
games).
Are there any pros or cons to having this featured enabled which is off by
default?

Thanks,

JR

 

[chip]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

"JR" <JR@nowhere.net> wrote in message news:ZOykd.2887$4U1.2836@trndny05...
>I just enabled Native Command Queuing on my 74GB WD Raptor . I was curious
>if anyone with the same setup has seen any improvements (particularly with
>games).
> Are there any pros or cons to having this featured enabled which is off by
> default?

I can't see it causing any problems - since the controller and the disk both
support it.

But on the other hand all it does - as I understand it - is optimises the
disk seeks being done in a heavy multi-tasking environment, e.g. in a server
with lots of users simultaneously requesting different data at different
positions on the disk.

This being the case, I doubt you'll see much benefit on a single user home
PC.

Chip

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

On Thu, 11 Nov 2004 00:58:33 GMT, "JR" <JR@nowhere.net> wrote:

>I just enabled Native Command Queuing on my 74GB WD Raptor . I was curious
>if anyone with the same setup has seen any improvements (particularly with
>games).
>Are there any pros or cons to having this featured enabled which is off by
>default?

Actually I had command queuing enabled with my Raptor on my K8N Neo2 board and
the performance was terrible; the Read Burst speed as tested from the Device
Manager properties dialog was 5.0 (this is with the 446 driver and BIOS 1.3)
and the sustained speed was about the same. Disabling this got my speeds back
to normal ranges (108 and 71.4). Other utilities also confirmed this poor
performance (e.g. HDTach)... in fact, I first noticed the problem when I
started playing games like EverQuest 2 and virtual memory swapping was
dropping my frame rates to about 1-2 fps.

I think there is a bug somewhere in the BIOS or drivers as theoretically
having this feature on should improve performance (albeit primarily in
multi-access scenarios), not hurt it.

-Tom

 

[chip]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

"The Mad Downloader" <cablemodem@cablemodem.co.jp> wrote in message
news:j8jnq0pd7ht0834i1ucjpfjr4bbkh236s2@4ax.com...
> On Thu, 11 Nov 2004 00:58:33 GMT, "JR" <JR@nowhere.net> wrote:
>
>>I just enabled Native Command Queuing on my 74GB WD Raptor . I was curious
>>if anyone with the same setup has seen any improvements (particularly with
>>games).
>>Are there any pros or cons to having this featured enabled which is off by
>>default?
>
> Actually I had command queuing enabled with my Raptor on my K8N Neo2 board
> and
> the performance was terrible; the Read Burst speed as tested from the
> Device
> Manager properties dialog was 5.0 (this is with the 446 driver and BIOS
> 1.3)
> and the sustained speed was about the same. Disabling this got my speeds
> back
> to normal ranges (108 and 71.4). Other utilities also confirmed this poor
> performance (e.g. HDTach)... in fact, I first noticed the problem when I
> started playing games like EverQuest 2 and virtual memory swapping was
> dropping my frame rates to about 1-2 fps.
>
> I think there is a bug somewhere in the BIOS or drivers as theoretically
> having this feature on should improve performance (albeit primarily in
> multi-access scenarios), not hurt it.
>

I have the same board and disks. How did you enable (or disable) TCQ? (or
NCQ.... whatever)

Chip

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

On Tue, 30 Nov 2004 09:06:24 -0000, "Chip" <anneonymouse@virgin.net> wrote:

>I have the same board and disks. How did you enable (or disable) TCQ? (or
>NCQ.... whatever)

From Device Manager, select "IDE/ATAPI controllers" and do a
right-click/Properties on the nForce3 250 Serial ATA Controller of choice. My
Raptor drive is on the Primary Channel of the second controller. From the
appropriate Channel tab you can change the option for Command Queuing (reboot
required, I think) and run the speed test.

If you can, let us know what you find out (i.e. if it does or doesn't have the
speed issue I encountered).

 

[chip]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

"The Mad Downloader" <cablemodem@cablemodem.co.jp> wrote in message
news:nuipq0p5ad3laf94kgec3j7t9oel3vhd7p@4ax.com...
> On Tue, 30 Nov 2004 09:06:24 -0000, "Chip" <anneonymouse@virgin.net>
> wrote:
>
>>I have the same board and disks. How did you enable (or disable) TCQ?
>>(or
>>NCQ.... whatever)
>
> From Device Manager, select "IDE/ATAPI controllers" and do a
> right-click/Properties on the nForce3 250 Serial ATA Controller of choice.
> My
> Raptor drive is on the Primary Channel of the second controller. From the
> appropriate Channel tab you can change the option for Command Queuing
> (reboot
> required, I think) and run the speed test.
>
> If you can, let us know what you find out (i.e. if it does or doesn't have
> the
> speed issue I encountered).

Interesting.

I have not installed the nVidia IDE drivers - I am using the Microsoft
Windows XP2 ones - and the option is not there. I have had bad experiences
with nVidia IDE drivers before (CD burner issues, drives shutting down and
starting up again, all sorts of problems) so I thought I would not bother
installing them.

Everything seems to be working well. I get 140MB/s ATTO scores for both
reads and writes and hdtach shows an average 129.9MB/s read across the whole
array (Max 145MB/s, Min 106MB/s), with 237MBs burst and 8% CPU. This seems
OK for my drives on an nforce3 controller.

So I am reluctant to install the nVidia drivers and mess things up. I will
probably risk it, however.

But before I do, please can you do me a favour and (in Device Manager) check
what version of nVidia IDE driver you are running? I assume its the latest?
But even so, I am not sure what the latest IDE driver version is.

Thanks

Chip

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

Hi,

I have the K8N Neo2 board and have one 36gig rapter on the primary sata
channel and two 'NCQ' capable sata seagates on channels 3 & 4 in a raid0
config. The command queuing option on all these drives just is not
available. Under 'Properties' etc, the 'Enable Command Queuing' is there
but is greyed out.

Thanks,
Shane
"The Mad Downloader" <cablemodem@cablemodem.co.jp> wrote in message
news:nuipq0p5ad3laf94kgec3j7t9oel3vhd7p@4ax.com...
> On Tue, 30 Nov 2004 09:06:24 -0000, "Chip" <anneonymouse@virgin.net>
> wrote:
>
>>I have the same board and disks. How did you enable (or disable) TCQ?
>>(or
>>NCQ.... whatever)
>
> From Device Manager, select "IDE/ATAPI controllers" and do a
> right-click/Properties on the nForce3 250 Serial ATA Controller of choice.
> My
> Raptor drive is on the Primary Channel of the second controller. From the
> appropriate Channel tab you can change the option for Command Queuing
> (reboot
> required, I think) and run the speed test.
>
> If you can, let us know what you find out (i.e. if it does or doesn't have
> the
> speed issue I encountered).
>

 

[Spajky]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

On Wed, 1 Dec 2004 14:36:23 +1000, "no spam" <nospam@hotmail.com>
wrote:

>I have the K8N Neo2 board and have one 36gig rapter on the primary sata
>channel and two 'NCQ' capable sata seagates on channels 3 & 4 in a raid0
>config.

can I ask you to make me a favor?
I would like you to make me a bench of your Raptor & your drives on
Raid0 & tell me the index you´ve got with this older Sandra to
compare:

Real life average performance: use this Sandra to get Index:
http://users.volja.net/jerman55/SiSoftSandraMin.zip 459kB
read Readme file inside
Comparing indexes between different drives clearly shows their
practical difference in speed.

WD Raptor 76Gb 10k Rpm - gets index points of 41.500
mine QuantumF+AS 20Gb /7,2k Rpm - points 19.500

this means that mine 3y old one is more than once slower than that
Raptor!

TIA !
--
Regards, SPAJKY ®
& visit my site @ http://www.spajky.vze.com
"Tualatin OC-ed / BX-Slot1 / inaudible setup!"
E-mail AntiSpam: remove ##

 

[chip]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

"Spajky" <Spajky@##volja.net> wrote in message
news:d8nsq0l2gi191fkt4pb5gokbmkfe6ma3os@4ax.com...
> On Wed, 1 Dec 2004 14:36:23 +1000, "no spam" <nospam@hotmail.com>
> wrote:
>
>>I have the K8N Neo2 board and have one 36gig rapter on the primary sata
>>channel and two 'NCQ' capable sata seagates on channels 3 & 4 in a raid0
>>config.
>
> can I ask you to make me a favor?
> I would like you to make me a bench of your Raptor & your drives on
> Raid0 & tell me the index you´ve got with this older Sandra to
> compare:
>
> Real life average performance: use this Sandra to get Index:
> http://users.volja.net/jerman55/SiSoftSandraMin.zip 459kB
> read Readme file inside
> Comparing indexes between different drives clearly shows their
> practical difference in speed.
>
> WD Raptor 76Gb 10k Rpm - gets index points of 41.500
> mine QuantumF+AS 20Gb /7,2k Rpm - points 19.500
>
> this means that mine 3y old one is more than once slower than that
> Raptor!
>
> TIA !

OK, I got 81,000

I suspect it should be higher than that actually. The reason I say that is
because it seemed low to me so I ran Sandra2005 straight afterwards and that
scored 108, whereas it usually scores 120+. I suspect my disk fragmentation
is reducing the score a bit.

I guess mid 80's would be possible on a clean disk.

Chip

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

Spajky,

The drive index for the 36gig Raptor is 29178. The index for the Raid (2 X
Seagates) is 64747.

Regards,
Shane
"Spajky" <Spajky@##volja.net> wrote in message
news:d8nsq0l2gi191fkt4pb5gokbmkfe6ma3os@4ax.com...
> On Wed, 1 Dec 2004 14:36:23 +1000, "no spam" <nospam@hotmail.com>
> wrote:
>
>>I have the K8N Neo2 board and have one 36gig rapter on the primary sata
>>channel and two 'NCQ' capable sata seagates on channels 3 & 4 in a raid0
>>config.
>
> can I ask you to make me a favor?
> I would like you to make me a bench of your Raptor & your drives on
> Raid0 & tell me the index you´ve got with this older Sandra to
> compare:
>
> Real life average performance: use this Sandra to get Index:
> http://users.volja.net/jerman55/SiSoftSandraMin.zip 459kB
> read Readme file inside
> Comparing indexes between different drives clearly shows their
> practical difference in speed.
>
> WD Raptor 76Gb 10k Rpm - gets index points of 41.500
> mine QuantumF+AS 20Gb /7,2k Rpm - points 19.500
>
> this means that mine 3y old one is more than once slower than that
> Raptor!
>
> TIA !
> --
> Regards, SPAJKY ®
> & visit my site @ http://www.spajky.vze.com
> "Tualatin OC-ed / BX-Slot1 / inaudible setup!"
> E-mail AntiSpam: remove ##

 

[Spajky]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

On Thu, 2 Dec 2004 13:02:37 +1000, "no spam" <nospam@hotmail.com>
wrote:

>The drive index for the 36gig Raptor is 29178. The index for the Raid (2 X
>Seagates) is 64747.

Shane, thanks!

khm, I thought that 36G Raptor would have better result & Raid
Seagates worse that came out ...

You did it with that mine version of Sandra?
--
Regards, SPAJKY ®
& visit my site @ http://www.spajky.vze.com
"Tualatin OC-ed / BX-Slot1 / inaudible setup!"
E-mail AntiSpam: remove ##

 

[Spajky]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

On Thu, 2 Dec 2004 09:57:38 -0000, "Chip" <anneonymouse@virgin.net>
wrote:

>> Real life average performance: use this Sandra to get Index:
>> http://users.volja.net/jerman55/SiSoftSandraMin.zip 459kB
>> read Readme file inside
>> Comparing indexes between different drives clearly shows their
>> practical difference in speed.

>OK, I got 81,000

>I suspect it should be higher than that actually. The reason I say that is
>because it seemed low to me so I ran Sandra2005 straight afterwards and that
>scored 108, whereas it usually scores 120+. ....

I can not use Your result, since you have done it with latest Sandra .

That version (2001pro) that I proposed fopr HB benchmarking,
is the last one which uses test file not bigger than 0,5Gb & do not
depends on ammount of Ram & CPU installed; so all kinds of setups can
be directly compared. (even 5 or more years old)

With higher versions of Sandra thats not so & so you get with same HD
on faster/more powerful setups better results (not really comparable).

Sandra "grabs" all available Ram while doing this bench ...
Chipset/HD controller drivers (DMA also) has VERY important impact on
result ...

BTW, the result you´ve got, was it for 2x Raptor76G in Raid0 ?

--
Regards, SPAJKY ®
& visit my site @ http://www.spajky.vze.com
"Tualatin OC-ed / BX-Slot1 / inaudible setup!"
E-mail AntiSpam: remove ##

 

[chip]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

"Spajky" <Spajky@##volja.net> wrote in message
news:n174r0pbpv44r2035vvljm49nod4inr3ej@4ax.com...
> On Thu, 2 Dec 2004 09:57:38 -0000, "Chip" <anneonymouse@virgin.net>
> wrote:
>
>>> Real life average performance: use this Sandra to get Index:
>>> http://users.volja.net/jerman55/SiSoftSandraMin.zip 459kB
>>> read Readme file inside
>>> Comparing indexes between different drives clearly shows their
>>> practical difference in speed.
>
>>OK, I got 81,000
>
>>I suspect it should be higher than that actually. The reason I say that
>>is
>>because it seemed low to me so I ran Sandra2005 straight afterwards and
>>that
>>scored 108, whereas it usually scores 120+. ....
>
> I can not use Your result, since you have done it with latest Sandra .

No, I haven't. I used the one you link to.

Chip

 

[chip]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

"Chip" <anneonymouse@virgin.net> wrote in message
news:31g8r3F3bbbnfU1@individual.net...
>
> "Spajky" <Spajky@##volja.net> wrote in message
> news:n174r0pbpv44r2035vvljm49nod4inr3ej@4ax.com...
>> On Thu, 2 Dec 2004 09:57:38 -0000, "Chip" <anneonymouse@virgin.net>
>> wrote:
>>
>>>> Real life average performance: use this Sandra to get Index:
>>>> http://users.volja.net/jerman55/SiSoftSandraMin.zip 459kB
>>>> read Readme file inside
>>>> Comparing indexes between different drives clearly shows their
>>>> practical difference in speed.
>>
>>>OK, I got 81,000
>>
>>>I suspect it should be higher than that actually. The reason I say that
>>>is
>>>because it seemed low to me so I ran Sandra2005 straight afterwards and
>>>that
>>>scored 108, whereas it usually scores 120+. ....
>>
>> I can not use Your result, since you have done it with latest Sandra .
>
> No, I haven't. I used the one you link to.
>
> Chip

EDIT: Yes, that's with 2x74GB Raptors in Raid 0

81,000 with Sandra 2001Pro.

120,000 with Sandra 2005

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.msi-microstar (More info?)

Spajky,

> You did it with that mine version of Sandra?

Yes, used your version of Sandra.

Shane