Sign in with
Sign up | Sign in
Your question

RAID 0

Last response: in Storage
Share
November 20, 2010 7:11:47 PM

Hello!
Please help this noob :-P
I'm running RAID 0 with two Seagate ST3500841AS
http://img94.imageshack.us/img94/8932/clipboard01fp.jpg
But before the RAID 0, with only one Seagate ST3500841AS it was faster, better stats, etc...
Where did I go wrong?

More about : raid

a b G Storage
November 20, 2010 7:51:35 PM

I have two of those and that graph is better than a single drive.
A single drive wil have a max transfer speed of around 140MBs and an av of 104MBs. Min around 80MBs.
Your graph may be distorted slightly by the big dips. These may have been caused by the hard disk doing something else at the time.

A single Barracuda 7200.12 500GB will not give a max read of 200MBs. Nor a average of 177MBs.
Your RAID could go faster at max and I've seen graphs showing a max of around 250-260MBs. It may be limited by the controller.
m
0
l
a b G Storage
November 20, 2010 8:04:06 PM

Here one that was used for Overclockers.com,
http://www.overclockers.com/forums/showthread.php?t=641...

Here is how single drives normally perform.
http://www.****/reviews/123-Seagate-Barracuda-720012-500GB-Drive-Review-ST3500410AS-Page-3.aspx

Are you using a controller card or onboard RAID controller, which one?
m
0
l
Related resources
a c 99 G Storage
November 20, 2010 8:23:53 PM

Couple notations:

1. How do you know the single drive was "faster, better stats, etc..." Do you have a graph for the single drive?

2. You do realize that hard drives NEVER EVEN COME CLOSE to the 3.0Gbps bandwidth for SATA II. Even 2 in RAID 0 won't come close. You'd need 4 in RAID 0 with "drive striping" to get close.

3. If you want truely outstanding "benchmarks," get an Solid State Drive, for use as a boot drive.

4. Don' t be to hung up about one single benchmark stat. Run several. Still, benchmarks don't show "real time/real world usage."

5. IMHO, it's the NVidia chipset RAID controller at fault.
m
0
l
November 20, 2010 8:44:10 PM

1. Before the RAID 0 I used HD Tune on the single HD, I remember 227 average.

2. I will try with 4 later.

5. Probably I should buy a New Motherboard.
m
0
l
November 20, 2010 9:04:58 PM

foscooter said:
Couple notatopns:

1. How do you know the single drive was "faster, better stats, etc..." Do you have a graph for the single drive?

2. You do realize that hard drives NEVER EVEN COME CLOSE to the 3.0Gbps bandwidth for SATA II. Even 2 in RAID 0 won't come close. You'd need 4 in RAID 0 with "drive striping" to get close.

3. If you want truely outstanding "benchmarks," get an Solid State Drive, for use as a boot drive.

4. Don' t be to hung up about one single benchmark stat. Run several. Still, benchmarks don't show "real time/real world usage."

5. IMHO, it's the NVidia chipset RAID controller at fault.



Why is the Burst Rate too low?
I see other people posting too much higher.
m
0
l
a b G Storage
November 20, 2010 10:00:55 PM

I'm not sure what hapenned with the burst rate. It does not have a real effect on performance. It is meant to be the inteface limit. Your performance is higher than the burst rate witch is odd.

I would just like to say that a single one of you drives would not provide better performance than your RAID shows.

Upgrading your MOBO migh increase the maximum read but the minimum would stay the same.
m
0
l
November 21, 2010 4:11:41 AM

Plumble said:
I'm not sure what hapenned with the burst rate. It does not have a real effect on performance. It is meant to be the inteface limit. Your performance is higher than the burst rate witch is odd.

I would just like to say that a single one of you drives would not provide better performance than your RAID shows.

Upgrading your MOBO migh increase the maximum read but the minimum would stay the same.




• NV RAID supports 4 SATA II ports(SATA 1 to 4). Transfer rate is up to 300MB/s.
• Silicon Image's SATA RAID supports another 2 SATA II ports. (SATA 5 and 6)
- Transfer rate is up to 300MB/s.

NV Raid and Silicon both would do the same performance or one could do better than other?
m
0
l
a b G Storage
November 21, 2010 4:57:35 AM

foscooter said:
Couple notatopns:

1. How do you know the single drive was "faster, better stats, etc..." Do you have a graph for the single drive?

2. You do realize that hard drives NEVER EVEN COME CLOSE to the 3.0Gbps bandwidth for SATA II. Even 2 in RAID 0 won't come close. You'd need 4 in RAID 0 with "drive striping" to get close.

3. If you want truely outstanding "benchmarks," get an Solid State Drive, for use as a boot drive.

4. Don' t be to hung up about one single benchmark stat. Run several. Still, benchmarks don't show "real time/real world usage."

5. IMHO, it's the NVidia chipset RAID controller at fault.



Yes, I agree it could be the NVidia chipset RAID controller - hardware error. For me you cannot enjoy the speed of a RAID 0 having only 2 try 4 of them instead.
m
0
l
a b G Storage
November 21, 2010 2:47:53 PM

Is this your system drive? Remember, the drives you see tested are tested as secondary drives, with nothing on them, under the best circumstances. If your OS is on these drives, the readings you see will be lower because of the overhead and other things going on with drives while the test is running, hence the large spikes.
m
0
l
a c 127 G Storage
November 21, 2010 6:57:06 PM

Try benchmarking with CrystalDiskMark or AS SSD; HDtune is not really suitable to test RAID arrays and sometimes gives plain wrong (too low) results. CDM and AS SSD test on the filesystem instead; how other apps would be reading and writing too. This is the only relevant scores in my opinion. HDTune is synthetic because of the way it does I/O; raw single queue I/O while circumventing the filesystem.
m
0
l
!