• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Matrix RAID, perfect solution for home use?

T

tempoct

Weaksauce
Joined
Sep 16, 2005
Messages
120
Hi,
I'm building my new system (core 2 duo based) and selecting a storage system. I have nevered used RAID before and would like to try it out.
The Matrix RAID seems to be perfect for home use. It allows user to setup RAID 0 and RAID 1 on two drives. Seems interesting. Only bad thing I found (other than lose some diskspace) is risk on RAID 0 part, which should not host any critical data anyway.
http://techreport.com/reviews/2005q1/matrix-raid/index.x?pg=1
I had my mind setup for RAID 5 (HW controller) but it would needs 4 drive to be good and controller is $$. So I look at RAID 10/ 0+1 which is not required too much power from controller (southbridge based controller should be fine) but it is also required 4 drives minimum (noise problem).
Anyone tried the Matrix? Needs real world opinion since it looks good on paper.
tempoct
 
If you're not accessing both portions of the disk at once, it's fine. But putting your OS on one portion and games on the other is probably a recipe for disaster. Seeking all the way across the disk would destroy any performance advantage you could get from raid 0.

The standard gaming rig suggestion is a Raptor as a boot/games disk, and a large disk (320, 400, 500 gigs, take your pick) as a documents/movies/junk/backups disk. Add an external backup unit if you have stuff that's important to you.

 
Check this out....

http://www.xtremesystems.org/forums/showthread.php?t=120184

I was thinking of giving this a shot with my RAID0'd WD740ADFD's. Making 1 20GB RAID0 slice on the fastest part of the HDDs for the OS/pagefile etc. And creating a RAID0 with what's left over.

I can see your argument about seeking all the way across the HDDs. I don''t know if this guy has takin' that into account or not.

2ms better access time seems like it might be worth it...
 
Brahmzy said:
Check this out....

http://www.xtremesystems.org/forums/showthread.php?t=120184
Click to expand...

That's called "short-stroking", and I don't see anything particularly innovative about it. All that's going on from what I see, is that a fake drive is made which uses the outer-most sectors. Now, because this is an itty-bitty fake drive using the fastest part of several fast drives, the average performance looks great, as does the seek time, because the worse parts of the drive aren't used, and there isn't much to distance to seek across.

My point is that any drives would do this, regardless of "matrix" raid, if only the same part of the drive was occupied with data. I.e. a partition would work just as well. The difference between a partition and matrix raid here is that HD tach would include the measurements for the rest of the space of the drive, ignoring the partition information. But this is just an HD Tach issue, not something that would affect your overall performance any differently were you actually just using the same part of the drives.

The gist of this is not using the slower parts of the drives, thereby having a faster and closer used section. So you really get this performance only if you don't use the drives beyond this slow part. If you do, then you're sunk again, because of the larger seek times and the reduced data rate off the slower parts.

A classic size vs. speed trade-off, I'd say.
 
Madwand said:
That's called "short-stroking", and I don't see anything particularly innovative about it. All that's going on from what I see, is that a fake drive is made which uses the outer-most sectors. Now, because this is an itty-bitty fake drive using the fastest part of several fast drives, the average performance looks great, as does the seek time, because the worse parts of the drive aren't used, and there isn't much to distance to seek across.

My point is that any drives would do this, regardless of "matrix" raid, if only the same part of the drive was occupied with data. I.e. a partition would work just as well. The difference between a partition and matrix raid here is that HD tach would include the measurements for the rest of the space of the drive, ignoring the partition information. But this is just an HD Tach issue, not something that would affect your overall performance any differently were you actually just using the same part of the drives.

The gist of this is not using the slower parts of the drives, thereby having a faster and closer used section. So you really get this performance only if you don't use the drives beyond this slow part. If you do, then you're sunk again, because of the larger seek times and the reduced data rate off the slower parts.

A classic size vs. speed trade-off, I'd say.
Click to expand...

I think you're right. There's no difference in doing this vs. just a small boot partition. I'd rather do the partition as it can be resized. A Matrix volume cannot be.

This is only beneficial if you are indeed running different RAID types.
 
I think the Matrix RAID 0 also use the outermost tracks of the drive too. But you should have (in theory) double read rate since it read the "stripe" from two different physical drives concurrently.
 
You must log in or register to reply here.
Back
Top