Who knows about HDD benchmarking?

[W1zzard]

I am looking into adding HDD reviews to TPU ...

Does anybody here have experience how to properly benchmark drives? Please no Sandra/ATTO/HDTach unless you can explain how (cached?) read bandwidth can relate to single-/multiuser application performance in real life.

[Batou1986]

Sounds like a great idea as i am planing on getting a new hd soon maybe someone could make a test bench program that would load certain applications and report disk access info and over all read and write times.

heres one for APK

[W1zzard]

maybe someone could make a test bench program that would load certain applications and report disk access info and over all read and write times.

thats pretty much what IPEAK does .. trial version: http://sunsite.rediris.es/pub/mirror/intel/ipeak/ -> spt-v30.zip
trial version can be run only 5 times, then delete HKEY_CLASSES_ROOT\CLSID\{7CB54E10-146E-11D3-9C57-0004AC59AE06} and similar keys around there with the same NSPK stuff in it :)

lots of info about it on storagereview.com

[DRDNA]

Sounds like fun , but most hard drives are just plain ole SLOW , its been our biggest bottle neck !!!! only raptors and perpendicular drives are any good and in all honesty they are a bottle neck too...We would have to run raid 5 in order to over come the bottleneck...Its like WTH I need to buy at least 4 hard drives to overcome a bottle neck that in all reality shouldn't even be a bottle neck , the damn industry has been almost the same in performance since well for too many years as far as i am concerned!!!!Please MR.Manufacture give me a single hard drive that does NOT bottleneck my PC!!!!

[Batou1986]

with the new ssd drives (solid state disk) coming out soon we should over come that bottle neck

now come on ppl someone has got to know more about hardrives and a good way we could compare performance so throw some ideas out there

[Steevo]

Just simply create a database of drives copying a file of X size on X controller on XP clean install. Use a accurate way of timing each copy, and try disk to disk, RAID to RAID, Game level load, and Office-application load.


There will always be factors such as fragmentation that affect the drives speed, but with a clean install you can minimise the effects.



I also hit up Maxtor about testing drives awhile back, mentioned this site and aureal density in conjunction with read and write speed vs RPM. Our own HDD Tach thread shows conclusively that a higher aureal density drive will outperform in read and write speeds a higher RPM drive.


MS also has a database query tool for SQL that simulates network traffic that can test a drive system.




Just my .02 though.

[pt]

looks like a good idea, but i don't know any programm that could do that
can't you get a full version of IPEAK?

[Batou1986]

that could be aranged

[W1zzard]

databases are usually memory bound. also database and file copying is not what most people here are using their drives for. what ARE you using your drive for?

[Steevo]

I meant more of Drive X on Controller X should have this average performance.


My drives are busy with loading games, downloading files then moving around to make more sense out of the hundreds of gigs of crap I have. Not to mention transferring photos, editing music, video, and compressing video.


I still have a few DVD's of backed up media and stuff that I have to copy on, but I have been thinking about a new motherboard so it would be a waste of time.

[Steevo]

And our database at work, one of two are about 6Gb each, and are more disk bound. RAID 5 600GB of Maxtor Sata drives.


20 users for the databases, besides file sharing, incoming file recipient from our scanners, two F@H clients, and two users. So really disk performance was at the forefront. A gig of ram manages to make if decently fast for users, and the gigabyte connection and tweaked stacks make it good for file serving.


9355 results out of a few hundred thousand possibilities in 22 seconds.

[DaMulta]

Loading times in games would be very nice to see on HDD Vs. HDD

[Steevo]

Anyway, I'm just about done defragging and will be off to play more.



And, uh, I peed in your pool.



Yay.

[W1zzard]

Loading times in games would be very nice to see on HDD Vs. HDD

farcry, fear and world of warcraft come to mind with long loading times .. anything else?

[Wile E]

farcry, fear and world of warcraft come to mind with long loading times .. anything else?

BF2?

[W1zzard]

ah very good one

[DanTheBanjoman]

Besides performance I would say temperatures and noise are important as well. Specially when used in a home server, running a drive at 70C 24/7 is a bad idea.

[DaMulta]

Quake 4 can have some long loading times also.

[Steevo]

Internal and external data transfer are also important.



Say a 1, 5, 10GB file transfered from one folder to another then to another drive just like it.

[W1zzard]

temperature is hard to measure because depending on manufacturer the same spot has different heat coming on .. same for the smart temperature .. we could measure power draw though

[lemonadesoda]

How about an external probe to test temp of the HDD case? You could then be consistent between makes/models.

[W1zzard]

How about an external probe to test temp of the HDD case? You could then be consistent between makes/models.

where would you put the probe on the case? the case does not have a uniform temperature

[Urlyin]

Well I think whatever is used would also be available to the readers in regards to them being able to test their own drives for comparison... Which kind of puts you back in the Sandra, Everest type apps .... Have you asked your Everest friends?


I ran across the heat temp issue on the TT Aquabay M4 water cooler (http://www.techpowerup.com/reviews/Thermaltake/M4/6)...

[Steevo]

I am looking into adding HDD reviews to TPU ...

Does anybody here have experience how to properly benchmark drives? Please no Sandra/ATTO/HDTach unless you can explain how (cached?) read bandwidth can relate to single-/multiuser application performance in real life.

HDTach does use burst speed, and that in itself is somewhat meaningless, unless you are checking to be sure of what mode the drive is in, and possibly only to ATA drives as SATA has of yet to actually break the 3.0Gbps speed barrier.

But the disk sequential read is uncached, there are too big of write files on the "long test" to cache. CPU usage is somewhat unreliable though, as it seems to vary by as much as 2-3% per test and it even stated +/- 2%. But 2% these days isn't alot to most.

The sequential drive read speed is the most accurate way of telling about the true world performance of the drive-drive subsystem. It should directly relate to how fast the drive system is when loading game levels, as well as other disk intensive tasks.



Seek times seem important to users, however I seriously doubt the role they play in loading a game level, the data is sequential and laid out so on the disk if defragmenting is done. And as most small files are cached (windows startup files) when a system is optomised, and thumbnails are cached by windows, I see minimal difference in seek times here.



Bu to each their own and I hope that we can get conclusive results from testing done.




I could submit a few results off the different systems I admin, as well as my home system.


Here is our server, live on the network. I was-am the only active user besides network load, and it was marginal at the time.

[funkflix]

Hi,

i would prefer h2benchw from the german pc magazine c't - ftp://ftp.heise.de/pub/ct/ctsi/h2benchw.zip. It supports write tests too, but only on a unpartitioned drive and finally give u a "Anwendungsindex", which u can compare to all other HDDs tested. Only problem is, that the programm is in german, but i think we/you can handle that. :)

For example a quick(about 20 minutes) test on my WD Raptor:

Platte: WDC WD740ADFD-60NLR1
Kapazität: CHS=(9039/255/63), 145211535 Sektoren = 70904 MByte

Interface-Transferrate mit Blockgröße 128 Sektoren bei 0.0% der Kapazität:
Sequenzielle Leserate Medium (ungebremst): 81363 KByte/s
Sequenzielle Leserate Read-Ahead (Verzögerung: 0.87 ms): 111183 KByte/s
Wiederholtes sequenzielles Lesen ("Coretest"): 121829 KByte/s

Dauertransferrate (Blockgröße: 128 Sektoren):
Lesen: Mittel 70209.8, Min 44042.4, Max 82187.3 [KByte/s]

Zugriffszeit Lesen: Mittel 8.57, Min 0.05, Max 16.95 [ms]
Zugriffszeit Lesen (<504 MByte): Mittel 5.41, Min 0.05, Max 14.42 [ms]

Anwendungsprofil `Swappen': 16102.4 KByte/s
Anwendungsprofil `Installieren': 23639.0 KByte/s
Anwendungsprofil `Word': 31332.2 KByte/s
Anwendungsprofil `Photoshop': 25614.1 KByte/s
Anwendungsprofil `Kopieren': 44260.9 KByte/s
Anwendungsprofil `F-Prot': 14391.0 KByte/s
Gesamtergebnis: Anwendungsindex = 23.5
!!! WARNUNG: Anwendungsprofile ungenau, da nur lesend gemessen
ATA-Platte: WDC WD740ADFD-60NLR1
Seriennr. : WD-WMANS1241524
Firmware : 20.07P20
Standard-Version: ATA/ATAPI-7
Puffergröße: 16384 KByte
Unterstützte UDMA-Modi: 0 1 2 3 4 5 6
Kein UDMA-Modus aktiviert!
Kapazität (28-Bit-Adressierung): 145226112 Sektoren (70911.2 MByte)
Kapazität (48-Bit-Adressierung): 145226112 Sektoren (70911.2 MByte)
Akustik-Management nicht unterstützt.

Only read tests.

Maybe this tool is what u searched for? :)