6_6_6

NCQ: Best Upgrade For a Power User!

Recommended Posts

You still have not understood what is most important for a workstation (in contrary to a server).

In the very first posting where you answered my "problem" you already showed that you didn't understand, and you still don't.

The most important parameter is RESPONSE TIME!

Meaning: when I do a click, I cannot wait for 30s until the application reacts.

What you are saying has no relevance, i am not testing database or fileserver performance, but i will explain this to you again. You have not used IOMeter, and you have no idea what is what there. Again you are talking for nothing.

In the IOmeter example above, AHCI-enabled drive can do 84% more disk operations than the one without AHCI. Its average response time is 83% better than the one without AHCI as evident by total throughput and total I/O's sec.. Maximum I/O refers to queue depth... and it is of no relevance here since one drive is not doing NCQ.

Without AHCI, this drive only performs a single disk operation, and everything else waits until that disk operation finishes. With AHCI, drive access is reordered, and all applications requesting drive access get their drive time without waiting for the other.

This all is evident by all kind of drive tests i ran... hdspeed, dd, iometer. More throughput in simultaneous operations, snappy system.

Exactly. What I found in this thread seemed to be the help I needed (Seagate 7200.11 with working NCQ).

And since you praise them so much, it was only fair to report that it does NOT work for me, so that others that are considering to buy Seagates have a second and different point of view.

Post screenshots so that we can examine the results.

1. The Perc solves the "freezing application"-problem, because it behaves much better than the ICH9R (see above for details).

Post hdspeed and iometer screenshots of your Perc. You don't have to unwrap anything.

Share this post


Link to post
Share on other sites
@6_6_6 : obviously FAT_Punisher got a problem with his original controller,

That's why he bought an other one. I've never experimented the bug he described and probably you too,

so it's useless to discuss if the controller he bought was justified or not. This have nothing to do with NCQ at all.

Indeed, there was a problem.

xerces8 describes it in this thread as well, so I am not the only one, fortunately. ;)

We (xerces8 and me) are not sure if it's related to NCQ or not. The fact that it doesn't happen with the ICH9R in IDE mode supports that theory, because NCQ is disabled in IDE mode.

Since that problem was solved by the new controller, it's more than justified for me. ;) Because you can't create a RAID in IDE mode.

Regarding NCQ: when I'll have the money, I'll simply buy a 3ware which does it's own "NCQ" (I mentioned it somewhere). That seems to work even with WD discs, and it should be independent from the OS.

Share this post


Link to post
Share on other sites

@FAT_Punisher

Indeed, there was a problem.

xerces8 describes it in this thread as well, so I am not the only one, fortunately.

6_6_6 have a ICH9R ( motherboard is GA-X38-DQ6 , see official website for specifications )

you had a ICH9R before buying a new controller

and I have a ICH9R (motherboard is GA-X48-DQ6 ).

But ok I've read the xerces8 thread

and seems that many hard drive perform worse with AHCI enabled and ICH9R.

Since you wanted absolutely to do a RAID0, you bought a new controller ---> OK

Now the problem you got with your western digital drive before getting the new controller, is despite "total throughput" remained the same,

they were "big periods" (1 second, 10s, 1 minute, it's random)

where all the throughput was reserved to "one access" and all other access where at zero.

Which meant, while copying a big file for instance, you couldn't do anything else.

Now with the new controller, you still have a shitty throughput, but at least the whole sytem does not freeze during big period

while copying big files ---> OK

Note: it might just be less expensive to just get Samsung hard drive , as they don't exhibit the same behavior ;)

Now 6_6_6, said it's normal to alternate between full and zero throughput for each hard drive access.

This is how it was intended to work. So I guess that for 6_6_6, it alternate very FAST. Hence total throughput remain the same,

but I think it is confirmed by the graph.

See 6_6_6 testing:

seagate_NCQ.png

Don't you see it alternates very fast between two threads, and that they are complementary ? (when the zero throughput for one thread , there's 100% for the other).

Now see xerces8 testing:

3ahcifm3.png

From what I've understood , he launched 3 threads.

There's huge gaps, but the graphs are again complementary.

Unfortunately when the third thread is launched,

all the throughput is assigned to it , and everything else is freezed.

I guess there's no way to copy a big file, and listen music at same times.

So the issue is no that throughput is alternating between each threads,

but does it alternate enough FAST, to do multiple things at same time.

when I'll have the money, I'll simply buy a 3ware which does it's own "NCQ" (I mentioned it somewhere). That seems to work even with WD discs, and it should be independent from the OS.

according to wikipedia :

For NCQ to be enabled, it must be supported and enabled in the SATA host bus adapter and in the hard drive itself. The appropriate driver must be loaded into the operating system to enable NCQ on the host bus adapter

In other word to have NCQ, it must be supported by hard drive, controller, and the operating system.

@6_6_6

You said in one post that your hard drive was running first in SATA- 1 mode , and that you didn't removed the jumper.

I might make a mistake , but I thought that NCQ was an exclusive feature of SATA II .

Could you download & install hdtune, and check in the info tab that it says that NCQ is enabled ?

Finally I must say I've read something weird: I found people having "multi tasking", working worse with their SCSI drives.

AFAIK , NCQ is a feature that try to imitate technologies already implemented for SCSI drives.

Edited by extrabigmehdi

Share this post


Link to post
Share on other sites

NCQ is SATA-I. My drive works in SATA-I.

In the pics, you can see BIOS Sata/Raid AHCI Mode: AHCI or Disable. In AHCI, NCQ supported; in DISABLED, it is not.

Share this post


Link to post
Share on other sites

@6_6_6

NCQ is SATA-I. My drive works in SATA-I.

Not all sata drive have NCQ support.

When Western Digital produced first SATA2 drives , they were still without NCQ support (see example).

Choosing AHCI mode in the bios, is not a warranty that NCQ is enabled.

So you should still check with Hdtune :lol:

Share this post


Link to post
Share on other sites
@6_6_6
NCQ is SATA-I. My drive works in SATA-I.

Choosing AHCI mode in the bios, is not a warranty that NCQ is enabled.

In mine, it is.

This is a recent SATA-II drive. And all recent SATA-II drives produced after 2006 have NCQ. Besides, Seagate was the first to implement native SATA. I don't know about that WD. But WD had PATA-to-SATA bridges for a long time while other drives supported native SATA.

Besides, jumpering it SATA-1 does not make NCQ go away. It just cuts the signaling speed to 150 MB/s for controllers that cannot support it. It doesn't make a SATA-II drive a SATA-I drive. I have no need for 300 MB/s. Default SATA-I is fine.

HDTune does not work reliably for me. I have NCQ status from ATA commands. hdat2, mhdd, victoria... all can issue ATA commands from DOS.

Share this post


Link to post
Share on other sites

I'm very curious if someone would be able to test this with one of the new WD VelociRaptor. I'm curious how multiple read processes on it compares to a Seagate 7200.11.

Share this post


Link to post
Share on other sites

@Atamido

I'm very curious if someone would be able to test this with one of the new WD VelociRaptor.

6 pages , and I have still no definitive answer too...

I've read reviews for velociraptor at newegg , and found exactely 3 reviews of people complaining that their velociraptor does not outperform enough their seagate , and hence it's not worth the price . But it's with more than 200 reviews, and 80% rated 5 eggs.

In the other hand there's much more negative reviews for seagate 7200.11 at newegg : often DOA , or death after few month. Some people think that seagate downgraded quality because of the fusion with Maxtor. Some think that seagate have bad quality control. Some other think that seagate have bad quality because they hard drives are produced in thailand ...

If you have the cash, why not buy one 7200.11 and post your results ? :lol:

I can't now, because it's not available were I live ...

Share this post


Link to post
Share on other sites

I just got a Seagate 7200.11 750GB and have made tests with it. I also have Samsung F1 750 GB and 1TB drives. Mainboard is Asus P5K-E with ICH9R, Win XP SP3 and Intel AHCI drivers.

I see a big difference between these drives when I test with HD_Speed. Both Samsung models drop to 7MB/sec (14MB total) with two instances of HD_Speed. The Seagate remains at 45-55MB (100-105MB total). And I do see/feel improvements with the Seagate, but I'll have to make more tests with real programs/usage situations.

One thing I noticed: I use Total Commander for file management and copying. If I use it for copying, it still makes the Seagate/XP react slowly. I'll have to make more tests and tweak TC's copy settings since it looks like copying via Windows explorer doesn't do this and it allows me to multitask much better when copying files from the Seagate or between Seagate's partitions.

Share this post


Link to post
Share on other sites

@DVB2100

thanks for your review.

Curious that the performance boost doesn't work with the software "Total Commander".

Is your system more snappy ? Do you have encountered some kind of "stuttering" (i.e it freeze at regulars intervals) ?

Any other user review is still welcome.

Share this post


Link to post
Share on other sites
In the other hand there's much more negative reviews for seagate 7200.11 at newegg : often DOA , or death after few month.

Those were probably all associated with the initial batch with the buggy firmware. That was more than a year ago. Seagate upgraded few firmwares i believe. I have the latest SD15 and i believe last 6 months or so this is shipping and no newer firmware.

DOA... might be many reasons. Your chances of getting a DOA drive is the same with any purchase. Not an issue. Return. (PS: I had many many drives of all makes and models over years and never had a DOA).

Dead after few months... I haven't seen any drive of mine fail after passing a day of stress testing. But i did return few drives coz they were defective. They worked fine, but they had awful graphs... and i am sure those would have failed if used for an extended time. So it is up to the user to follow an educated approach prior to commiting data to any drive. Sure drive can die all of a sudden, but i wouldn't be worried about that if your equipment is high quality. Besides, that can happen to any perfectly good drive too (PSU crash, etc).

Seagate has a noise issue. It is noisy and it makes those awful clicking sounds frequently. Samsungs are definitely better for that. But this is a moot point when a velociraptor is in the equation.

DVB,

Could you please post some screenshots? If possible both for the Seagate and Samsung drives (2 instances). Did you run the same tests on Seagate by turning off NCQ in BIOS?

You can leave aside TC for now and use explorer for your disk operations for testing purposes.

I just got a Seagate 7200.11 750GB and have made tests with it. I also have Samsung F1 750 GB and 1TB drives. Mainboard is Asus P5K-E with ICH9R, Win XP SP3 and Intel AHCI drivers.

I see a big difference between these drives when I test with HD_Speed. Both Samsung models drop to 7MB/sec (14MB total) with two instances of HD_Speed. The Seagate remains at 45-55MB (100-105MB total). And I do see/feel improvements with the Seagate, but I'll have to make more tests with real programs/usage situations.

One thing I noticed: I use Total Commander for file management and copying. If I use it for copying, it still makes the Seagate/XP react slowly. I'll have to make more tests and tweak TC's copy settings since it looks like copying via Windows explorer doesn't do this and it allows me to multitask much better when copying files from the Seagate or between Seagate's partitions.

Share this post


Link to post
Share on other sites
Do you have encountered some kind of "stuttering" (i.e it freeze at regulars intervals) ?

I don't know what you mean by stuttering or freeze at regular intervals but i never had a system that does that. Otherwise i would not be using it.

What i mean is that whenever you do any disk operation, you will have to wait until it is over for all the rest of the operations you want to initiate (that will have anything to do with disk). But freezing randomly at regular intervals??? I have no idea what that is.

Share this post


Link to post
Share on other sites

@6_6_6

But freezing randomly at regular intervals??? I have no idea what that is.

I saw many peaks in the graphs you showed (i.e it's almost "oscillating"), and was wondering how it would be translated in real world use, and if it could affect negatively user experience.

Share this post


Link to post
Share on other sites
@6_6_6
But freezing randomly at regular intervals??? I have no idea what that is.

I saw many peaks in the graphs you showed (i.e it's almost "oscillating"), and was wondering how it would be translated in real world use, and if it could affect negatively user experience.

Oh, now i understand. Nothing of this sort.

Share this post


Link to post
Share on other sites

@6_6_6

Besides, that can happen to any perfectly good drive too (PSU crash, etc).

Well, that's why I see how many reviews are negative at newegg, which would give an idea of "failure rates". I guess that's not always reliable , but that still helps.

I would be quite angry if my PSU crashed.

Seagate has a noise issue. It is noisy and it makes those awful clicking sounds frequently. Samsungs are definitely better for that.

I find very curious that many people said that they found their seagate relatively quiet.

I know that seagate doesn't have AAM , but they've compensated this with a new special motor "SoftSonic". So I was assuming that noise was not necessarily a big issue.

I really hated my previous Seagate because of the constant "clicks", and was thinking it was defective because of this. From what you are saying I'd rather command a Scythe Quiet Drive with my seagate. But the hard drive might be overheating then.... <_<

But this is a moot point when a velociraptor is in the equation.

Why do you say it 's a moot point ? For lot of people velociraptor is quiet. Some people are complaining of "high pitched" sound, but it seems it doesn't affect everyone. The velociraptor, has even been recommended by "silent pc review", which means it can be very silent. Problem: "silent pc review" advise to remove heatsink... (which amplify noises).

Edited by extrabigmehdi

Share this post


Link to post
Share on other sites
I find very curious that many people said that they found their seagate relatively quiet.

I do have a 0 db system. I have been running silent systems over 10 years. People might say many things since hearing degradation is gradual and permanent. Everyone's hearing threshold is not the same. People also do have 20 fans in their systems and they think it is quiet. Seagate was producing quiet drives and were favorable back then with Barracuda IV and V and that is what we have been using back then. They are no longer up to par, and Samsung's are better for this job for a long time coming. I had both Samsung's and Seagate and my Samsung is whisper quiet compared to this one.

Why do you say it 's a moot point ? For lot of people velociraptor is quiet. Some people are complaining of "high pitched" sound, but it seems it doesn't affect everyone.

I don't find it quiet. It exactly has that high-pitched sound similiar to SCSIs. They are not loud, but it is not quiet either. But then again, my 15K Cheetah was not a whole lot more noiser than my 200GB Caviar back then. But it just was not quiet. When you have other sources of noise, it does not matter much. But when you have nothing, that single source amplifies to the front.

Share this post


Link to post
Share on other sites

@6_6_6

But when you have nothing, that single source amplifies to the front.

You are right, I got a psu (antec truepower quattro) which drived me crazy because of high pitched sound . It was definitely not loud, but other part of my rig were much quiter.

Now, I'm trying to replace it, just to ensure it was not a defective unit. Lot of people say under review that this psu is quiet, which is a bit embarrassing ...

Edited by extrabigmehdi

Share this post


Link to post
Share on other sites
Curious that the performance boost doesn't work with the software "Total Commander".

Is your system more snappy ? Do you have encountered some kind of "stuttering" (i.e it freeze at regulars intervals) ?

The Total Commander has many options for file copy: selectable block sizes for big/small files and also different block size if the file is copied to the same disk or another disk, it also has two copy modes and also via Explorer method. So I think I will find good copying performance when I have time to check this.

System does feed snappier, looks like XP's bootup is also faster. And if I copy files from the Seagate and try to open a video file from it, it doesn't take forever like it did with the 1TB Samsung. I have not encountered stuttering with the Samsung drives and had zero problems multitasking when eMule was downloading etc. Also, I had no problems when a TV recording program (DVBViewer) was recording at the same time as I did other things. Only problem was that during heavy disk access the system was blocked, i.e. the hard drive was not multitasking. But I'll have to make more tests to actually verify this in real situations, first impressions could be wrong...

Share this post


Link to post
Share on other sites
Could you please post some screenshots? If possible both for the Seagate and Samsung drives (2 instances). Did you run the same tests on Seagate by turning off NCQ in BIOS?

I'll do the tests and post screenshots later today. I have already removed the F1 750GB drive, but it was identical to the 1TB drive. I have also a Spinpoint 500GB Samsung that I can test. So I will test the Seagate, 1TB F1 and 500GB Samsung (HD501LJ). I haven't tested the Seagate without NCQ, I'll do also that with all these drives.

Share this post


Link to post
Share on other sites
The Total Commander has many options for file copy: selectable block sizes for big/small files and also different block size if the file is copied to the same disk or another disk, it also has two copy modes and also via Explorer method. So I think I will find good copying performance when I have time to check this.

DVB2100,

Please leave the Total Commander aside now and use only Windows Explorer for testing purposes. It is better not to deal with something else to troubleshoot with. Once you are comfortable, of course you can go back to TC and make it work the way you want it to. But at this point, it would add one more layer to the equation that we need to deal with. Not everyone is using TC. Vanilla install is explorer. Don't you think so?

System does feed snappier, looks like XP's bootup is also faster. And if I copy files from the Seagate and try to open a video file from it, it doesn't take forever like it did with the 1TB Samsung.

Yes, that was exactly my impression too. I did have the Samsung 500GB T166.

I don't think you would get much perf improvement with booting, and it is not that important anyway. How many times do we reboot our PCs?

It is not important for us, but if you want to time your bootup. Here is a small VBS script to make your life easier:

Option Explicit
On Error Resume Next
Dim Wsh, Time1, Time2, Result, PathFile, MsgResult, MsgA, AppName, KeyA, KeyB, TimeDiff
MsgA = "Please close all running applications and click on OK."
KeyA = "HKEY_CURRENT_USER\Software\RestartTime\"
KeyB = "HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Run\RestartTime"
AppName = "ReBoot-Time"
Set Wsh = CreateObject("WScript.Shell")
PathFile = """" & WScript.ScriptFullName & """"
Result = wsh.RegRead(KeyA & "Times")
if Result = "" then
MsgResult = Msgbox (MsgA, vbOKCancel, AppName)
If MsgResult = vbcancel then WScript.Quit
Wsh.RegWrite KeyA & "Times", left(Time,8), "REG_SZ"
Wsh.RegWrite KeyB, PathFile, "REG_SZ"
Wsh.Run "cmd /c Shutdown -r -t 00", false, 0 
else
Wsh.RegDelete KeyA & "Times"
Wsh.RegDelete KeyA
Wsh.RegDelete KeyB
TimeDiff = DateDiff("s",Result,left(Time,8))
MsgBox "Your computer reboots in " & TimeDiff & " seconds", VbInformation, AppName
end if
wscript.Quit

Just save this to a text file, rename the text file to: time_reboot.vbs

and double-click it. It will initiate a reboot and show you how long it took when it comes back up. (make sure you are not shutting down the pc and adding removing hardware at this point).

It will delete registry keys after it shows you the time.

Thank you for your contribution.

PS: I don't remember the source, i am not the author.

Share this post


Link to post
Share on other sites
Could you please post some screenshots? If possible both for the Seagate and Samsung drives (2 instances). Did you run the same tests on Seagate by turning off NCQ in BIOS?

I'll do the tests and post screenshots later today. I have already removed the F1 750GB drive, but it was identical to the 1TB drive. I have also a Spinpoint 500GB Samsung that I can test. So I will test the Seagate, 1TB F1 and 500GB Samsung (HD501LJ). I haven't tested the Seagate without NCQ, I'll do also that with all these drives.

I tested HD501LJ in this first post in this thread. I don't think people are interested in that drive any way. You can leave that aside if you wish. 1TB F1 would be beneficial to everyone.

How are you measuring the disk throughput? Does XP have similiar tool to Vista's TASK MANAGER > PERFORMANCE > RESOURCE MONITOR?

Are you satisfied with your Seagate?

Share this post


Link to post
Share on other sites

In Windows XP : Control Panel / Administrative tools / Performace

This will get you the performance monitor that is also available in Vista (I think there is a button that starts it in "Resource Monitor"; if not, it can be started over the control panel, like in XP).

Once running , click the plus button and add PhysicalDisk / "Disk Bytes/sec"

(sorry if this has already been posted...)

Share this post


Link to post
Share on other sites
Please leave the Total Commander aside now and use only Windows Explorer for testing purposes. It is better not to deal with something else to troubleshoot with. Once you are comfortable, of course you can go back to TC and make it work the way you want it to. But at this point, it would add one more layer to the equation that we need to deal with. Not everyone is using TC. Vanilla install is explorer. Don't you think so?

I don't think you would get much perf improvement with booting, and it is not that important anyway. How many times do we reboot our PCs?

Yes, will use just Explorer to test performance during file copy etc. What I noticed is that it looked like Explorer lowered it's copy process priority when I did other things, but TC just kept copying full speed making the system react slow (but it didn't take CPU time and lowering priority did nothing). I was disappointed about that and thought that despite the HD_Speed test results, NCQ does not really work. But then I tried copying with the Explorer and was surprised how fast everything (Word, Excel, Photoshop) loaded. But I will have to verify this today, it was late so maybe I was just dreaming...

I reboot my PC once per day so speed it's not a major issue. It's just something I noticed, I have few scripts running during bootup, I also load Firefox and virus program. And there definately was something different/faster in the way those load now with the Seagate.

Share this post


Link to post
Share on other sites
@6_6_6
But freezing randomly at regular intervals??? I have no idea what that is.

I saw many peaks in the graphs you showed (i.e it's almost "oscillating"), and was wondering how it would be translated in real world use, and if it could affect negatively user experience.

Oh, now i understand. Nothing of this sort.

By the way, when graph is oscillating, the disk activity is not paused or at a complete halt on the system to cause a stutter. On the contrary, as can be seen in the throughput, the disk is still pushing 105 MB/s just like it does on a single instance. It is just that the disk/OS is reordering the requests and pushing something to each application instead of waiting the previous one to finish in order to start the rest.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now