• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

which to use: on-board RAID or a Windows striped volume

NTB

Diamond Member
Wasn't sure whether to post this here or in one of the HW forums...anyway: I have a pair of SATA drives that I'd like to chain together to use as 'scratch' space - fast, temporary storage, mostly for video and image editing. But, having never tried this before, I'm wondering which would be the better way to go:

(1) Use the RAID capabilities built into the motherboard (GIGABYTE GA-EP45-UD3R)

-or-

(2) leave the disks as normal drives and set up a striped volume via Windows Vista's 'computer management' control panel.

anybody have any suggestions?

And, in case anybody's curious, some basic specs:

GA-EP45-UD3R main board

Intel Q9400 CPU (Core2 Quad, 4x2.66GHz, 6MB L2 cache)

6GB DDR2 RAM

64-bit Vista Home Premium

Thanks!

Nathan
 
Originally posted by: Nothinman
Use hardware raid, it will be much faster.

No, it really won't.

So if a dedicated chip on MB does all the work instead you propose to let windows and computer CPU do it and you still say it will not be faster? I did had configured a few raid systems even raid 0+1 and while I could do it with windows I prefer hardware raid just because it has better speed, better failsafe and health monitoring options, does not interfere with whatever software you will install on the raid itself.

 
So if a dedicated chip on MB does all the work instead you propose to let windows and computer CPU do it and you still say it will not be faster?

With RAID0, yes I would say that. But it's irrelevant in this instance anyway because those cheapo onboard RAID controllers don't have any real hardware in them. All of the work is done in the driver on the host CPU so you end up with the host CPU doing the work plus you're at the mercy of probably crappy 3rd party drivers.

I did had configured a few raid systems even raid 0+1 and while I could do it with windows I prefer hardware raid just because it has better speed, better failsafe and health monitoring options, does not interfere with whatever software you will install on the raid itself.

Neither RAID0 or RAID1 require any real processing, hell even software RAID5 isn't a problem with CPU speeds where they are these days. So speed isn't an issue. The way it handles a failed drive is irrelevant in this instance because the OP is talking about RAID0 which is a dead array no matter what, health monitoring will probably actually be better with software RAID because you can use any old tool to monitor the SMART status of each drive, if the OS just sees an array it either won't be able to read the SMART data or it'll be bogus. And I've never heard of software interfering with a software RAID array so I have no idea what you mean by that.
 
i prefer hardware RAID as well

If you mean a real hardware RAID controller with battery-backed cache and such, then sure. But if you just mean one of those cheapo, onboard fakeRAID controllers then, no, there's no reason to like them over software RAID. Infact software RAID is much more flexible and the performance will at least be the same if not better.
 
Originally posted by: NTB
anybody have any suggestions?

Try them both, running some simple benchmarks, e.g. an ATTO 256 MB sweep with Direct I/O = "Neither", e.g. copying a large set of files to one directory and then making a copy of that directory on the same array, observing and timing the transfers. Or try a test of some processing you'd actually do on the machine which would be I/O, not compute bound.

If you don't see much of a difference, go with Windows RAID for greater portability. If you don't want to bother measuring, go with Intel RAID -- it's known to perform well. Windows RAID 0 should in theory perform just fine, but Windows RAID has been sad in other cases, so I wouldn't bet on it without measuring first.
 
Originally posted by: Nothinman
So if a dedicated chip on MB does all the work instead you propose to let windows and computer CPU do it and you still say it will not be faster?

With RAID0, yes I would say that. But it's irrelevant in this instance anyway because those cheapo onboard RAID controllers don't have any real hardware in them. All of the work is done in the driver on the host CPU so you end up with the host CPU doing the work plus you're at the mercy of probably crappy 3rd party drivers.

I did had configured a few raid systems even raid 0+1 and while I could do it with windows I prefer hardware raid just because it has better speed, better failsafe and health monitoring options, does not interfere with whatever software you will install on the raid itself.

Neither RAID0 or RAID1 require any real processing, hell even software RAID5 isn't a problem with CPU speeds where they are these days. So speed isn't an issue. The way it handles a failed drive is irrelevant in this instance because the OP is talking about RAID0 which is a dead array no matter what, health monitoring will probably actually be better with software RAID because you can use any old tool to monitor the SMART status of each drive, if the OS just sees an array it either won't be able to read the SMART data or it'll be bogus. And I've never heard of software interfering with a software RAID array so I have no idea what you mean by that.

Nothinman is entirely true.. Onboard RAID controllers are not 'hardware' raid controllers unless they have their own dedicated cpus.. and they almost always do not.

A true 'hardware' raid controller might be faster - but hardware raid cards are expensive.. usually $350+ (unless you buy a Perc5i pulled from a server...)..

Windows' built-in raid is pretty decent.. you can configure it on the fly without rebooting. I doubt the onboard raid implementation of cpu-dependent raid is any faster than Windows' built-in implementation..

Plus, Windows raid volumes can be moved to other Windows boxes*.

*disclaimer: I have not done this.. but it is supposed to work.
 
*disclaimer: I have not done this.. but it is supposed to work.

Yea, supposedly you can import foreign dynamic volumes although I haven't done it myself either. Most of my experience is with Linux software RAID which "just works".
 
Back
Top