Samsung SSD's in RAID 0

Printable View

Show 100 post(s) from this thread on one page

07-12-2008, 10:24 PM
I_G

Samsung SSD's in RAID 0

4 X Samsung 32GB SSD in Intel ICH9R RAID 0 write cache enabled:

http://img369.imageshack.us/img369/9...rixraidbe4.jpg

http://img113.imageshack.us/img113/5830/ssd634xm2.jpg

http://img104.imageshack.us/img104/1...ntelraiqi6.png
07-12-2008, 10:31 PM
Xcel

nice, what's the price of one of those?
07-12-2008, 10:44 PM
I_G

$399, $100 more than a Velociraptor.

http://www.newegg.com/Product/Produc...82E16820147053
07-12-2008, 10:55 PM
dinos22

that's really great

have you used it much for your OS
how does it feel like....can you see the responsiveness even in trivial tasks

what about some HDD intense programs....

great results so far
07-12-2008, 10:57 PM
1Tanker

Nice performance.:) I'm wondering why it starts out slower then picks up...then drops off at the end. SSD's usually have a somewhat even performance..from start to finish, and aren't slower at the ends like a magnetic drive. :confused:
07-12-2008, 11:29 PM
I_G

Quote:

Originally Posted by dinos22

that's really great

have you used it much for your OS
how does it feel like....can you see the responsiveness even in trivial tasks

what about some HDD intense programs....

great results so far

Yes, I use the Intel Matrix RAID so one array is my 24/7 work machine and the other array I use for benchmarking. I was previously using 6 150GB Raptors in RAID 0.

Most tasks are noticeably faster. The array defrags in about 20 - 30 seconds. The machine boots up about 12 -15% quicker.

The only thing I do that would be disk-intensive is image cloning, which if going from the RAID to a single SSD, or vice versa, it takes about 90 seconds to read and write a 60GB driver image.
07-12-2008, 11:50 PM
I_G

Quote:

Originally Posted by 1Tanker

Nice performance.:) I'm wondering why it starts out slower then picks up...then drops off at the end. SSD's usually have a somewhat even performance..from start to finish, and aren't slower at the ends like a magnetic drive. :confused:

At this point I'm guessing it's the RAID controller. The pattern is very consistent and turning the write cache on or off has no effect. I used to also get some strange patterns with the six Raptors.
07-12-2008, 11:57 PM
dinos22

Quote:

Originally Posted by I_G

Yes, I use the Intel Matrix RAID so one array is my 24/7 work machine and the other array I use for benchmarking. I was previously using 6 150GB Raptors in RAID 0.

Most tasks are noticeably faster. The array defrags in about 20 - 30 seconds. The machine boots up about 12 -15% quicker.

The only thing I do that would be disk-intensive is image cloning, which if going from the RAID to a single SSD, or vice versa, it takes about 90 seconds to read and write a 60GB driver image.

how does image cloning performance compare to normal drives
07-13-2008, 12:52 AM
eva2000

very nice.. hopefully next week i'll have 4x32GB OCZ SSD 120-140/90 read/write ones from local Aussie retailer

you plan to try with hardware raid controllers too or just onboard ICH9R ?
07-13-2008, 12:54 AM
Movieman

That's $1600.00 for an effective 128 gig drive.
I got to pass at this point but nice numbers!:up:
07-13-2008, 03:00 AM
Brother Esau

To Rich for my blood.
07-13-2008, 03:01 AM
WiZARD_hun

Quote:

Originally Posted by I_G

Most tasks are noticeably faster. The array defrags in about 20 - 30 seconds. The machine boots up about 12 -15% quicker.

defrag is the worst thing you can do with a SSD...
07-13-2008, 03:09 AM
1Tanker

Quote:

Originally Posted by WiZARD_hun

defrag is the worst thing you can do with a SSD...

I didn't think SSD's need defrag. :confused:
07-13-2008, 03:15 AM
tiro_uspsss

Quote:

Originally Posted by WiZARD_hun

defrag is the worst thing you can do with a SSD...

:confused: please explain? :yepp:

:up:
07-13-2008, 04:40 AM
bowman

The controller of an SSD deliberately 'fragments' the drive when writing to even out the wear cycles as much as possible, hindering that one chip dies with many other perfectly good chips still alive but unable to be used.

I guess in theory defragmenting it would force the controller to not do that, but I'd think it would be smart enough to not actually defrag the drive.
07-13-2008, 04:42 AM
tiro_uspsss

Quote:

Originally Posted by bowman

The controller of an SSD deliberately 'fragments' the drive when writing to even out the wear cycles as much as possible, hindering that one chip dies with many other perfectly good chips still alive but unable to be used.

I guess in theory defragmenting it would force the controller to not do that, but I'd think it would be smart enough to not actually defrag the drive.

wow! I never knew that! thats pretty darn interesting! thank u muchly for sharing! :toast: :wave:
07-13-2008, 06:59 AM
Xcel

I haven't read much about defragmenting SSDs but OCZ for instance doesn't recommend it on their new core SSDs.

Quote:

IMPORTANT NOTE: Solid State Drives DO NOT require defragmentation.It may decrease the lifespan of the drive.
07-13-2008, 08:00 AM
eva2000

Quote:

Originally Posted by Movieman

That's $1600.00 for an effective 128 gig drive.
I got to pass at this point but nice numbers!:up:

how you calculate that.. in OZ 32GB OCZ Core SSD 2.5" retails for AUD$219/each so = AUD$876 (~US$850) and 64GB and 128GB OCZ SSD are AUD$339 and $599 respectively http://i4memory.com/f59/ocz-core-ssd...ustralia-7785/ :)

Puts the 64GB version around same price as 300GB WD raptor for Australian prices.
07-13-2008, 09:00 AM
Xcel

I'm tempted to go 2x32GB OCZ Core for my OS :D (available for preorder here 172€ each)
07-13-2008, 10:35 AM
bowman

I'm waiting it out till Intel's 80GB drives this fall - perhaps something to accompany the Nehalem freshness. :)

http://www.hothardware.com/News/Inte...SD_Sneak_Peek/

Quote:

Also in the system were a pair of 80GB Intel-built solid state drives running in a RAID 0 configuration. After seeing these drives in action and talking with representatives from Intel, we now know why Intel is excited to bring these drives to market. The combination of an Intel designed SATA II controller with high speed flash has resulted in a wicked-fast SSD. Unfortunately, again, we were asked not to disclose benchmark scores, but we can say the sustained transfer speeds were significantly higher than anything else we have seen to date.
07-13-2008, 02:53 PM
I_G

Quote:

Originally Posted by 1Tanker

I didn't think SSD's need defrag. :confused:

They don't, but I just had to see how fast it would be.

And with a MTBF of 228 years I figured once or twice wouldn't hurt...
07-13-2008, 03:00 PM
I_G

Quote:

Originally Posted by dinos22

how does image cloning performance compare to normal drives

Four 150GB Raptor array - 60GB partition- 3+ minutes; 500GB partition 4+ minutes.

Six 32GB SSD's - 60GB partition - 45 seconds; 118GB partition - 90 seconds.
07-13-2008, 03:09 PM
dinos22

Quote:

Originally Posted by I_G

Four 150GB Raptor array - 60GB partition- 3+ minutes; 500GB partition 4+ minutes.

Six 32GB SSD's - 60GB partition - 45 seconds; 118GB partition - 90 seconds.

wow i'd become religious if they could invent some cool SSD religion :D

this is very cool indeed
i'll definitely get into some SSD action when i do my next update on 24/7 system
07-22-2008, 11:59 PM
xtclocker

Wow, that's is really fast!! beautiful drives
07-23-2008, 12:04 AM
gillll

any difference between 2 and 4 drives ?

Show 100 post(s) from this thread on one page

All times are GMT -8. The time now is 09:55 AM.

XtremeSystems