[cluberti]

Check your PM.

[gunsmokingman]

Got it and sent a reply

Other Test Results

Hd Test 2 Old Computer


HD Test 2 New Computer


I still do not know if my HD are using the 1.5 or 3.0 interface. I sent a Email to HP and waiting
for a reply.

This post has been edited by gunsmokingman: 10 February 2009 - 02:43 PM

[cluberti]

Look, the tests you post are testing sequential read speeds with burst rate, and drive seek time (not actual time to do anything once the seek is done). They don't test any caching, random read and write, or anything else that would stress the firmware. Those tests are cosmetic at best - they mean nothing compared to the Winsat tests, which go deeper in testing the cache performance, read and write random speed on multiple size files, rather than just how fast the seek is or how fast you can sequentially stream a file (which should be fast on almost any drive).

Again, I respectfully disagree - your testing methods are not thorough for accurately testing drive performance. They're a part of what a drive can do, but definitely not something that would take into account any of the above methods Winsat performs. Also note that you have older drives that are slower at these superficial tests that perform better - meaning they likely have better firmware and can handle the odd loads that the real world would give them better.

I would seriously recommend opening an administrative command prompt and run winsat disk -v on both old and new disks, and see exactly where the discrepancy is. It's not random, I assure you - those older disks really are doing better at the winsat tests than the Seagate 500s.

For example, I have a 250GB WD SATAII drive that scores a 5.5, and a winsat disk -v run shows this:

C:\Windows\system32>winsat disk -v
Windows System Assessment Tool
...
> Disk  Sequential 64.0 Read                51.89 MB/s   5.4
> Disk  Random 16.0 Read                     1.39 MB/s   3.4
> Average Read Time with Sequential Writes  11.715 ms    2.9
> Latency: 95th Percentile                  36.290 ms	 2.5
> Latency: Maximum                          89.616 ms    7.7
> Average Read Time with Random Writes      13.741 ms    2.8
> Total Run Time 00:00:50.87

[puntoMX]

cluberti, on Feb 10 2009, 03:46 PM, said:

... - meaning they likely have better firmware and can handle the odd loads that the real world would give them better.

Indeed, that must be the bottle neck in this case...

EDIT: By the way, that's an ICH9R / G33 chipset, so the firmware of the RAID/ACHI controller must screwing up your score. I wonder what HP has t say about this, IF they even reply as you are using a OS that isn't supported by them.

Here's mine:

C:\Windows\system32>winsat disk -v
Windows System Assessment Tool
...
> Disk  Sequential 64.0 Read				   191.74 MB/s		  7.3
> Disk  Random 16.0 Read					   2.68 MB/s			4.3
> Average Read Time with Sequential Writes	 8.030 ms			 4.9
> Latency: 95th Percentile					 14.594 ms			5.1
> Latency: Maximum							 21.969 ms			7.9
> Average Read Time with Random Writes		 7.502 ms			 5.2
> Total Run Time 00:00:55.05

[cluberti]

puntoMX, on Feb 10 2009, 04:52 PM, said:

EDIT: By the way, that's an ICH9R / G33 chipset, so the firmware of the RAID/ACHI controller must screwing up your score. I wonder what HP has t say about this, IF they even reply as you are using a OS that isn't supported by them.

Note that a vast majority of folks using this specific drive model score 3.0 maximum, across a whole host of systems, so I doubt the controller is the root cause. Unless they're all using a ICH9R/G33's, which would be quite coincidental .

[gunsmokingman]

Ran the tests on both computers
Uploaded the results in a zip file.
Responce from HP about the interface speed in Raid

Quote

I have verified the product specifications and found that the 1 TB (2 x 500) SATA 3G (3.0 Gb/sec) hard drive was bundled with your PC.

I will also run the same test on Vista Sp2 and post the results.

This post has been edited by gunsmokingman: 18 February 2009 - 10:11 PM

[cluberti]

                                               WD 400            Seagate 500
Disk  Sequential 64.0 Read                  62.06 MB/s   5.7   153.70 MB/s   7.1
Disk  Random 16.0 Read                       1.01 MB/s   3.0     2.41 MB/s   4.2
Average Read Time with Sequential Writes     5.991 ms    5.7     8.462 ms    4.7
Latency: 95th Percentile                    10.676 ms    5.6    45.524 ms    1.9
Latency: Maximum                            83.275 ms    7.7    94.381 ms    7.7
Average Read Time with Random Writes         6.379 ms    5.5    14.475 ms    2.7

Note the horrid blocking latency averages and average read with random write times compared to the two. You get great sequential performance (athough sequential write isn't even as fast as the WD 400, which is interesting), and decent random reads of larger files, but when it comes time to actually do a lot of work that would hit the cache and stress the firmware (the 95th percentile latency number registers how fast or slow blocking-induced latency is at the 95th percentile of the mean vs the maximum latency average number, and average reads with random writes hitting the disk is stressing the write caching and read-method portions of the firmware controller, specifically), the drive falls flat on it's face. Sorry to say, but Seagate built a drive tuned for streaming larger files around, not for random access. Sorry, the tests don't lie.

[gunsmokingman]

Ok if i accept your explaination of what happening, then explain why on Vista I get 5.9 versus 2.8 on the 2 computers.
Now the disk scores are as follows New 5.9 versus Old 5.6, from what I understand is that I should get a lower score
on Vista with my set up. Which is not the case I get a higher disk score on Vista with the Seagate then the WDC, so is there a
problem with Vista that it produces the wrong results.

[cluberti]

gunsmokingman, on Feb 10 2009, 06:40 PM, said:

Ok if i accept your explaination of what happening, then explain why on Vista I get 5.9 versus 2.8 on the 2 computers.
Now the disk scores are as follows New 5.9 versus Old 5.6, from what I understand is that I should get a lower score
on Vista with my set up. Which is not the case I get a higher disk score on Vista with the Seagate then the WDC, so is there a
problem with Vista that it produces the wrong results.

Vista doesn't do any latency checks like this, only the more superficial sequential tests. You can run winsat from the command line in Vista as well and see the differences. The Vista score is *not* actually accurate with regards to disks, and this was rectified in Win7's WinSat. Unfortunately, there are going to be a lot of people like you who feel like this is wrong, but the numbers do not lie (and Seagate is generally known for making drives like this, this is not a new thing). If you want help looking at the winsat scores from Vista and Win7 to get an idea of what they mean, let me know.

[gunsmokingman]

HP sent me a fix for Windows7 low access score.

Quote

From: pavilion_support_en@mail.support.hp.com [mailto:pavilion_support_en@mail.support.hp.com] 
Sent: February-10-09 3:48 PM
To: Gunsmokingman@TELUS.NET
Subject: Re: RE: RE: HP Pavilion Elite m9360f Desktop PC e-

 

Hello Ed, 

Thank you for writing back. 

Ed, In reviewing your email, I understand that an error message low index score.

As per the PC specification the PC is bundled with the SATA 2.0 and it have 3.0 Gb/sec speed. With the information you have provided to us you have installed Windows 7 in your PC. 
 
We at HP do not have the tested information regarding the Windows 7. 

However, I am providing the troubleshooting steps which may resolve the issue.

Fix Low Hard drive Windows Experience Index Score/Rating on Windows 7
1. Right Click on My Computer and Select Properties from context menu.

2. Select Device manager from left pan

3. Locate your Primary Disk Drive and right click select Properties

4. Now Uncheck Enable Write Caching on Drive and Press ok

5. re-run your Windows Experience Index

Please visit the below weblink which has graphical representation of the troubleshooting steps. So, please perform the troubleshooting steps form the Weblink.

http://www.blogsdna.com/2133/fix-low-hard-drive-windows-experience-index-wei-score-in-windows-7.htm

NOTE:  The URL above will take you to a non HP Web site.  HP
			does not control and is not responsible for information
			outside of the HP Web site.

I did what was suggestted and got a 6 on the index, now I am happy
Could you post the cmd line for winsat on vista I tried winsat disk -v but got a error.

So I guess I was wrong to some extend, the fix was so simple I had to hit myself with
those wet noodles.

[cluberti]

This is what I told you to do January 26th . Note that you should re-run the Win7 winsat tests just to see the difference.

As to Vista, you can still run winsat disk, but you can't use the -v option (it doesn't support it for the disk tests).

As you can see, the Seagate drive's write caching performance is poor, so keep that in mind when you shop for your next hard disk .

[gunsmokingman]

When you posted that about the disk i was still running this in non raid,
didnt occour to me it could affect the raid set up.
As to the hard drives I wont be replacing them soon as I said I am happy
with the new results, I only expected a result of 5.5 to 6.5.

This is the new result of winsat disk -v

Quote

> Disk  Sequential 64.0 Read				   163.21 MB/s		  7.2
> Disk  Random 16.0 Read					   2.46 MB/s		  4.2
> Average Read Time with Sequential Writes	 8.276 ms		  4.8
> Latency: 95th Percentile					 13.101 ms		  5.3
> Latency: Maximum							 49.755 ms		  7.9
> Average Read Time with Random Writes		 9.364 ms		  4.1

This post has been edited by gunsmokingman: 10 February 2009 - 10:42 PM

[puntoMX]

There is also an option for Write back cache in the Intel Matrix software, enable it there and set the setting back for Windows 7 again.

* (If you didn't already played with the Matrix Software.)

[gunsmokingman]

The software does not install, but thanks for that information.

[cluberti]

The very latest version of the Intel Matrix software should work on Win7 (even x64), as I am currently running it. However, it failed to install the first Win7 box I built, but worked the second time. Odd, but I guess that's beta.

[DigeratiPrime]

Guys the reason Write Back Cache is disabled by default is it can greatly increase the chance of data loss. If you have data in cache and there is a power failure that data is lost. If this happens to data that is to be mirrored (RAID 1) the volume will be degraded, and if it is to be stripped (RAID 0,5) the volume will fail. The cache is only good for short bursts, because if there is a large amount to write the cache will fill and you will have to wait, at normal velocities, for it to commit to disk. If you look at discreet RAID controllers they recommend caching but only with a battery backup.

[gunsmokingman]

DigeratiPrime, thank you for the information about that. This is the first time I have used raid
on a computer.

[Access Denied]

I used Acronis to reload win7 on my RAID and it lowered my HD score from 5.8 on a single Raptor to 3.0 with a pair of them @ RAID 0, lol. Something ain't right with this scoring. The same drives on Vista went from 5.6 to 5.9 when RAID was enabled.

> Disk  Sequential 64.0 Read					   	   122.02 MB/s		  6.9
> Disk  Random 16.0 Read								  1.61 MB/s		  3.6
> Average Read Time with Sequential Writes				6.415 ms			   5.5
> Latency: 95th Percentile								 66.250 ms		   1.9
> Latency: Maximum									  383.346 ms		  5.3
> Average Read Time with Random Writes					  14.015 ms		  2.7

Some of these ms times are alot higher than his. Do they look right?


EDIT. I know the disk performance is better in RAID. From the way the pc runs, to benching the RAID @ 2x the speed of the drives alone.

This post has been edited by accessdenied042: 15 February 2009 - 07:41 PM

[cluberti]

accessdenied042, on Feb 15 2009, 08:30 PM, said:

I used Acronis to reload win7 on my RAID and it lowered my HD score from 5.8 on a single Raptor to 3.0 with a pair of them @ RAID 0, lol. Something ain't right with this scoring. The same drives on Vista went from 5.6 to 5.9 when RAID was enabled.

> Disk  Sequential 64.0 Read					   	   122.02 MB/s		  6.9
> Disk  Random 16.0 Read								  1.61 MB/s		  3.6
> Average Read Time with Sequential Writes				6.415 ms			   5.5
> Latency: 95th Percentile								 66.250 ms		   1.9
> Latency: Maximum									  383.346 ms		  5.3
> Average Read Time with Random Writes					  14.015 ms		  2.7

Some of these ms times are alot higher than his. Do they look right?


EDIT. I know the disk performance is better in RAID. From the way the pc runs, to benching the RAID @ 2x the speed of the drives alone.

If it's *specifically* 3.0, with latencies like that I'd suspect disabling the write cache will clear that up.

[Access Denied]

It is disabled and I cannot enable it(7 gives me an error stating it cannot be enabled) Only in RAID though, when just SATA, have all options there available, even the disable the write behind thingie. Im happy with the performance of it, even the score is wrong due to something. I know its faster, it works faster and that is all that really matters. I just thought the higher scores were too high for me to be running 74GB raptors in raid vs what he had. I know they are prob faster being sata II but I know for a fact my access times are lower per disk than his(raptors are bout half of reg drives, or used to be).