HDD testing

[Zak33]

Hi guys,

I just purchased two Western Digital 320 gig SATA drives.
WD3200AAKS (desktop drive, slightly cheaper)
http://www.wdc.com/en/products/Products.asp?DriveID=299


WD3200YS RE16 (enterprise edition, slightly more dough)
http://www.wdc.com/en/products/Products.asp?DriveID=233


I intend to see if the enterprise version is faster or slower in normal daily Hexus style use (ie games, film encoding, boot up, defraging etc)
I got them both to test against each other, on the same system, with fresh installs, identical in every respect.

I'm going to use XP Pro SP2, and HD Tachometer, and I shall use IL-2 1946 and Q4 for boot times.

What else should I use to benchmark them?

I'm open to offers. I'm installing XP on both tonight and will test tomorrow, so I'd like some ideas and links pls to free downloadable benchmarking software for hard disks.

Cheers, look forward to your help.

[Defenestration]

1) Copy lots of small files from same drive to same drive.

2) Copy large file(s) from same drive to same drive.

3) Copy lots of small files from one drive to other drive.

4) Copy large file(s) from one drive to other drive.

[aidanjt]

tests 1 and 2 are pointless if the copying is within the same partition, since all the Filesystem will do is rewrite the file descriptor tables.

[Defenestration]

tests 1 and 2 are pointless if the copying is within the same partition, since all the Filesystem will do is rewrite the file descriptor tables.

This is not the case with XP (FAT and NTFS) at least - the files will still actually be recreated and re-written.

[acrobat]

I used something called "HD Tach". It does a proper benchmark of any hard disk.

Just remember a few things though:

1) You are meant to boot up from fresh and then let your PC settle for 5 mins or so first.

2) Make sure absolutely nothing is running in the background that might read/write from the drive while you are testing it - and screw up your results.

3) The test results vary. So its worth testing each drive about 5 times or so. Then you can see what the drive achieves during those tests. Then test the other drive and it should hopefully show you which is faster.

[Zak33]

I've been doing all those, and it's very interesting.

Replicating large files on the same HDD is the most strenuous, without a doubt.

But I've also just sussed a faster boot time for XP.

[acrobat]

But I've also just sussed a faster boot time for XP.

Do share Mines pretty quick actually but I'm interested anyway.

Also, I forgot to mention, I got quite a nice program called "HD Tune" which benchmarks hard disks too. It also checks they are working ok aswell, and shows the temperature of them. Its a little bit reduntant though nowdays, since that new version of SpeedFan is showing me the temperature of everything short of my armpits and stuff. Its quite accurate now too - for me atleast. I quite liked the way HD Tune works though anyway.

[Zak33]

http://forums.hexus.net/showthread.php?t=108431

[Zak33]

OK: Results so far

Copying a 13.3 gig file from a Barracuda SATA to the Desktop model AAKS

3m 30 secs
3m 28 secs
3m 30 secs
3m 30 secs


Copying the 13.3 gig file from Barracuda SATA to Enterpise model YS RE
3m 58 secs
3m 57 secs
3m 56 secs
3m 58 secs

So consistently faster for the cheaper drive.

Copying a file from the same drive to the same drive, using the same 13.3 gig file

Desktop AAKS
9m 10 secs
9m 17 secs
9m 11 secs

Enterprise YS RE
8m 58 secs
8m 58 secs
8m 58 secs

Uber consistant and faster for the more money drive

HD tacho next
AAKS desktop drive

Burst Rate 205mb/s 204mb/s 205 mb/s
Rndm Access 13.2 ms 14.5 ms 13.3 ms
CPU use 5% 5% 5%
Average Read 69.8mb/s 69.5mb/s 69.8 mb/s

YS RE enterprise drive
Burst Rate 202mb/s 202mb/s 203mb/s
Rndm Access 14.5 ms 13.3 ms 13.8 ms
CPU use 5% 5% 5%
Average Read 57.1mb/s 56.7mb/s 57.0mb/s

Suprised about this, the average read speed is faster on the cheaper drive, as is the Random Access times


XP Boot times
AAKS Desktop drive
37 secs
36 secs
37 secs

YS RE enterprise drive
38 secs
38 secs
38 secs


So, in short, the cheaper drive is faster so far in most stuff

[Defenestration]

Those results are pretty much what I expected. The AAKS is the newer model with less platters, but the NCQ is slightly better on the YS RE model, which accounts for the better results on the same drive to same drive test.

[acrobat]

http://forums.hexus.net/showthread.php?t=108431

cool

[Zak33]

I'm keeping the Enterprise drive (slower though it is) as it'sgot a Molex.

And the other goes to my mate for his PC

[acrobat]

Could write time be a factor in why these drives are supposed to be good?

In other words, the read time may be a few seconds slower than some drives, but maybe they write stuff really quickly?

[Zak33]

In normal use it seems as fast as the AAKS.

[ati]

nice one zak33

it doesnt look like a huge amount of difference tbh.

what i dont get is why the aaks are not perp?

[Zak33]

perp? you mean Perpendicular?

I'm not 100% sure they'd need to be. That technology is clever, but not always fastest.