Has anyone tried to use Raid-0 to improve performance?

With the ubiquitous availability of SSDs, I don't think RAID-0 has the same appeal as it once did.

BRuys
With the ubiquitous availability of SSDs, I don't think RAID-0 has the same appeal as it once did.

From what I know about RAID-0 and the various tests that I have seen, apparently the speed is increased, linearly proportional to the number of units such SSD or HDD, is not it?

rogeriodec

BRuys
With the ubiquitous availability of SSDs, I don't think RAID-0 has the same appeal as it once did.

From what I know about RAID-0 and the various tests that I have seen, apparently the speed is increased, linearly proportional to the number of units such SSD or HDD, is not it?

Yes, it is, until you saturate the SATA port.  I have been building RAID arrays in servers for 20 years, usually with redundancy (RAID-1, 5, 6, 10, etc).  RAID-0 is fast, but with every additional disk, increases the chance of data loss, as one failed disk means you lose all your data.  If your motherboard supports RAID, great, but keep in mind that many 3rd party RAID chipsets do not properly support TRIM on SSDs.
 
If you need RAID-0, fine, but you would need to be recording many, many tracks to make it worth while in my opinion.

Doktor Avalanche
I mirror myself. I strongly recommend RAID 5 or RAID 1. RAID 0 does not have parity do not recommend. SSD is good but whatever you do backup...

RAID-5 suffers slow writes, due to the generation of parity data.  Professional level RAID controllers mitigate this somewhat with a large battery-backed write-cache, but even so, RAID-5 is not usually used with systems that do a large amount of write cycles.  I wouldn't use RAID-5 for a DAW.

That's a good idea Kev.

BRuys
With the ubiquitous availability of SSDs, I don't think RAID-0 has the same appeal as it once did.

Let's just focus on the subject of this topic is: speed, performance.
Obviously, a backup routine is essential. I myself have two automated daily backup processes, of the entire disk partitions, one to External HDD, other to Cloud.
Thus, the risk of data loss is the same using RAID-0 or not.
So, in any environment, backup is everything.
 
Now talking about performance. I read some benchmarks tests, I got so far to the following conclusions (all about SSD): 

1. Raid-0 is faster in almost all cases. According to this test, it is absurdly FASTER sequential transfers (large files) and slightly faster in random transfers. 
2. Still others argue that Raid-0 is best in any situation like this article.

However, when we use large sample libraries (I have almost 2 tb of samples), the price to store it all in SSD is, at least today, much more expensive than using HDD.
 
In short, in my opinion, if you use a lot of disk space and want more performance, Raid-0 in one (or more) HDDs is still the best and cheapest solution.

For huge sample libraries then HDD storage certainly makes economic sense. For less huge libraries SDDs have an advantage in seek time over HDDs in any configuration, so might work out more efficient. I don't use samples other than drums much so I've never needed to worry about them, the maximum I'm ever likely to need fit easily into the available RAM.

The many GB of photographs, videos and media files I've got are currently on HDD thanks to economics (though Photoshop uses SSD space as a scratch pad which does pay off).

One advantage SSDs have over HDD RAID arrays in DAWs or video editing machines is that SSDs are silent. They also draw less power and run much cooler than HDDs so a lower wattage fanless psu is viable and fewer case fans are needed to dump heat, again making things quieter.

Configuring an SSD setup is going to get interesting as drive prices drop. A multi-drive setup all on SSD should be very, very fast if load sharing is set up carefully. As Grem says, the SATA bus can be the limiting factor with current technology.

My question is how do you really benefit from the RAID 0 speed increase? Could your DAW not cope before?

Grem
Roger I am also wondering about the new "3D" SSD's that are just now coming out. It's predicted that these drives will eventually cause all SSD prices to fall. Their performance is better than the regular "2D" SSD's
 
For performance, it all revolves around the SATA bus. SSDs already saturating this bus, so a new bus has to be adapted before we can really see performance boosts more than we have now.
 
As far as sample libraries, I think Ram would have a bigger impact than the drive. Unless we are talking about "streaming" samples. Then I would say that SSD would benefit this process and would also get a bigger boost when the new "3D" SSD's become available along with the new bus.
 
 

Very intersting: http://fossbytes.com/inte...acity-at-lower-prices/

I agree with the good Doktor ...I don't see it being feasible to do that. I'd love to see a comprehensive head-to-head comparison for my own edification, but it wouldn't be worth paying for. I'd also like to know who does the serious heavy-lifting to make getting 12 ssds worth it.