Crapy Raid Performance?

[Shark-357]

Hmm i didnt know 36gb raptors were that slow in raid
Anyway here are mine hitachi 2x120 sata in raid 0 using 16K stripe

[MaxxxRacer]

I just went through the mess of convertion the drive to 16k blocks. Found out that my previous efforts had failed and that it was still 4k. This should improve read and write performance, but it didnt seem to demonstrate itself as such when I ran HD tach.

[Grinch]

go into device manager and make sure drives look like this...




I have found w/WD drives they like 16K and 32K stripe sizes...

[Shark-357]

Nice boost in burst speed thx for the tip Grinch

[soundx98]

It's awful long winded, but I just tested NV Raid and Sil3114 controllers on a DFI Venus. Tested Single drives, 2 and 3 drive RAID 0 arrays, SATA 1 Raptors vs SATA 2 16M cache WD2500KS, effects of various stripe/cluster combinations, NV vs Sil3114, effects of "Read Cache" Enabled/Disabled, the effects of "Command Queing" enabled/disabled, with screenshots (73) using HD Tach, ATTO, HD Tune, and Everest Disc Benchmark.

I'm certainly not an expert but you may find the screenshots very interesting as well as helpful. http://www.planetamd64.com/index.php?showtopic=21897

[Delirious]

go into device manager and make sure drives look like this...




I have found w/WD drives they like 16K and 32K stripe sizes...

I did this as it was recomended on dfi-street and did a before and after and noticed and increase in speed.

for what its worth, ATTO is a better benchmark than HDtach, i usually run it 3 times cause every time it can be significantly different.

Sata two only makes a difference on Huge raid 0 arrays (4+ drives) as well as with Burst speed.

[MaxxxRacer]

Grinch. thanks for the tip. I will re-bench the next time I restart to see if it helped.

[Haltech]

It's awful long winded, but I just tested NV Raid and Sil3114 controllers on a DFI Venus. Tested Single drives, 2 and 3 drive RAID 0 arrays, SATA 1 Raptors vs SATA 2 16M cache WD2500KS, effects of various stripe/cluster combinations, NV vs Sil3114, effects of "Read Cache" Enabled/Disabled, the effects of "Command Queing" enabled/disabled, with screenshots (73) using HD Tach, ATTO, HD Tune, and Everest Disc Benchmark.

I'm certainly not an expert but you may find the screenshots very interesting as well as helpful. http://www.planetamd64.com/index.php?showtopic=21897

Incase no one bothered to tell you, nice write up bro!

[soundx98]

Thanks man, teacher gave me a "T" for time consuming on it

[mesyn191]

Hmm i didnt know 36gb raptors were that slow in raid
Anyway here are mine hitachi 2x120 sata in raid 0 using 16K stripe

Its not that they're slow in RAID its that NVRAID sucks ass compared to a hardware RAID controller.

http://img242.imageshack.us/my.php?i...rmaxlineii.jpg

[scottk]

Ghost 10.05

Does 10.05 have these parameters?

-ffx -fni

I had to turn those on using Ghost 2003 to identify RAID arrays on the DFI.

[scottk]

The wierdest thing is that right now Im on the clone 30gb spare drive right now.. VERY wierd to do a straight copy.. it works fine, but I think it has some regisytry issues because Photoshop CS2 was complaining that it needed me to reactivate it.

I've had this happen with Windows XP activation too, when moving images between arrays and single drives.

[odb]

heres my 74gig raptors on my a8r32-mvp

[Grinch]

this is what I am getting on a 16K/16K raid 0 setup with my 2 wd 150 raptors on nvraid:

[Cooper]

Grinch what do you mean by 16K/16K - what second number stands for ?

[Hassan]

Also where other drives are catching up and surpassing the transfer rate of first gen raptors, their random access and seek times are still quite excellent

[safan80]

The only use SATA 2 has for raw performance is with the RAM drives.

anyway, while I would love 2x150gb raptors or even the 74g's its not worth the money to me. (not that I have it currently)

For setting up your stripe size you might want to check http://www.techspot.com/vb/topic1596.html

I may have discovered a formula for determining an appropriate stripe size.

If you go into XP disk defragmenter and click on Analyze for your array, it should return a value of the average file size.

Take this number, and divide it by 2 x # of harddrives in the array, then round DOWN to the nearest available stripe size you can choose.

Let's assume that the average file size on your harddrive is 512Kb.
Divide that by a factor of the above calculated number. In my case, I have 4 harddrives in my array, so I would choose 8.

So, 512Kb divided by 8 = 64K (the optimal stripe size for my average files).

Let's say you had 2 harddrives in the array, then using the above formula, you would choose 128K stripe size.

This is contingent on whether you are just storing data and seldom using it or are actively using the data. If just storing Data, then put the operating system on it's own harddrive, not on the array. OS=small stripe size, data=big stripe size in your case.

Another important OVERLOOKED factor is the Cluster Size of the formatted harddrives.

In my humble opinion, I would set the cluster size for 1/2 of the stripe size when formatting. This should reduce disk fragmentation and can appreciably decrease drive maintenance.

I have done several experiments, and smaller stripe sizes for the OS can make it really fly! But with smaller stripe sizes, you also have to defragment daily (not my cup of tea). Personally, I am using a 64K stripe with a 32K cluster. My fragmentation is pretty normal as opposed to a single harddrive. And I wouldn't suggest going with any stripe size less than 8K - you go from extreme speed at first to a snails pace in a few hours of heavy use, without defragmentating. Highpoint can let you go lower than that, and up to 2048Kb if desired.

It works quite nicely except on boot up, but that could be the umteen things I have running...

I would experiment further, but I wore out my XP disk and had to buy another one - lol. And Microsoft doesn't like it when you have to keep activating numerous times for your experiments.

To the guy with the MP3's and such, I suggest the following combination -

4Mb average filesize = 512 stripe size
Cluster size when formatting under XP should be 1/2 x 512 or 64K cluster size (max cluster size allowed under XP).


This should give you very nice performance and you shouldn't have to defragment too often.

By all means, email me. I would love to hear how well this works for you.

I was meaning the proformance of current sata2 based hard drives like the seagate 7200.9 series drives with 16MB cache in a raid0 setup and the not sata2 standard. However, speaking in terms of the standard the sata2.5 standard has hotplug and staggered spin up the latter which is useful when your dealing with a lot of drives. a good controller card is also worth getting because the nv controller doesn't have any onboard cache. I'm using an areca pci-express sata2 controller card in my other system and it is very fast to say the least.


you want to see large fast arrays go to http://forums.2cpu.com/ in the storage section.

[Hassan]

that is the purpose of diskeeper on a raid array, quick dfrags, cpl of mins, and an automated function to allow daily defrags at low usage times.

[MaxxxRacer]

well If I ever happen upon a good amount of spare cash, I will get one the areca cards (from what I've seen of that card it blows the socks off other PCIE STA cards) and run a 3x150gb raptor array. But until then I will stick with the onboard SATA.


and wow.. some guys is running 8x 150gb raptors in a raid array... thats absolutley insane..

EDIT: after that tweak the burst went to 204mb/s and the cpu utilization went in half.. an excelent tweak if u ask me.

[Delirious]

well If I ever happen upon a good amount of spare cash, I will get one the areca cards (from what I've seen of that card it blows the socks off other PCIE STA cards) and run a 3x150gb raptor array. But until then I will stick with the onboard SATA.


and wow.. some guys is running 8x 150gb raptors in a raid array... thats absolutley insane..

EDIT: after that tweak the burst went to 204mb/s and the cpu utilization went in half.. an excelent tweak if u ask me.

WHERE!!!

yah those areca cards are nice, especially the ones with onboard memory.

[Timmay]

go into device manager and make sure drives look like this...




I have found w/WD drives they like 16K and 32K stripe sizes...

In my device manager all the boxes are ticked... Should I untick the two that aren't in your screen shot?

[safan80]

In my device manager all the boxes are ticked... Should I untick the two that aren't in your screen shot?

No leave them alone.
leave write caching enabled
and bios select transfer mode enabled

[Timmay]

Well thats strange, My results are:
32K stripe, 8K cluster - 2 x Seagate 80GB's on nForce4 Chipset.
"4 Ticks": 220.8MB/s
"2 Ticks": 235.5MB/s

**edit**: Im about to ghost my setup and change to 64K stripe as I feel that works best for me.

[Grinch]

well If I ever happen upon a good amount of spare cash, I will get one the areca cards (from what I've seen of that card it blows the socks off other PCIE STA cards) and run a 3x150gb raptor array. But until then I will stick with the onboard SATA.


and wow.. some guys is running 8x 150gb raptors in a raid array... thats absolutley insane..

EDIT: after that tweak the burst went to 204mb/s and the cpu utilization went in half.. an excelent tweak if u ask me.

np...

[Grinch]

In my device manager all the boxes are ticked... Should I untick the two that aren't in your screen shot?

after doing a bunch of reading...the doing a bunch of tests this is what gave me best results on nforce4 chipset....there is a performance hit when NCQ is enabled as well as read caching...with read caching disabled it gives a more accurate picture...with NCQ disabled to get more speed out of it unless it is a multiuser environment....




here are some links:
http://www.storagereview.com/article...500ADFD_8.html
http://www.storagereview.com/article...500ADFD_4.html
http://www.storagereview.com/article...500ADFD_6.html
http://72.14.207.104/search?q=cache:...s&ct=clnk&cd=4
http://72.14.207.104/search?q=cache:...&ct=clnk&cd=17