10,000 RPM only on SCSI AND SATA 1.5?

Find some performance articles. There's not benefit from the extra bandwidth, but there is a benefit to the manufacturer to support a larger customer base (Non-tech geek type people don't know about capatability between different s-ata types).

So sata 3 wouldn't be faster, and sata 1.5 makes it more universal?


Stitches

That is correct.

People who love technology just for the sake of technology or the yearning for bragging rights will want something with SATA 3.0 and techies who love technology and actually know what they're buying and understand how it will benefit them, are likely to not care.

The only times during which SATA 3.0 will benefit even the fastest of today's SATA drives, is the very short burst moment when transferring data to and from the relatively tiny caches on those drives. The only time anyone will ever see a measurable difference is on benchmarking tools that measure peak burst transfer rates. Otherwise the SATA 1.5 spec is already capable of bandwidths far in excess of the maximum sustained transfer rates of today's SATA drives (1.5 Gbps in SATA 1.5 vs around .9 Gbps MSTR for a current WD Raptor).

-Brad

Despite is't popularity with enthusiasts, Western Digital is still trying to aim the Raptor at the Commercial market as a cheap alternative to SCSI.

I'm a musician and work with A LOT of plugins. I saw some gamers rant and rave that the WD 10,000 made their games load and run faster. Plus I gotta love the 16MB cache although I know I could get that with a 7200.

The main thing is that I am building a computer that will be upgradeable for atleast 3-4 years to come.

Because of this I am also going from AMD to intel, simply because intel has a 754 Quad, and I know in the future the board I buy now will accomadate the Quad core. Plus alot of board supporting 8 or more GBs of ram, and although it seems excessive, I have a dual core 64 bit AMD right now with 2GBs of RAM and sometimes it chockes, hard.



Stitches

WEAK! I just checked all my progs out and they're 180Gbs+ including the OS. Looks like I'm stickin to 7200 anyways.


I wish they had something larger. I know that the price is high, but this is for a professional recording studio and every ounce of performance is time, and time is money.


I could stack 2 150s, but Installing to 2 different drives makes plug ins lose their librarys, scripts, and just screw up all sorts of stuff.

I'm probablly moving to Vista ASAP. Is there any way to make XP or Vista think that 2 drives are really one large drive?




RAID duh. I'm losing my mind!

?? are you for real?!?

If time is money then get four 150GB Raptors and use RAID 0+1.

The stripe part will get you one big 300GB volume so you don't have to worry about the software not being able to find its plug-ins and libraries, and the mirror part, which will come at a ~ $450 price tag today for the two drives, will guarantee that your studio production does not screech to a halt when the inevitable hard drive failure occurs. Money well spent in my opinion.

If you really, really want to be on top of the performance curve, another $450 or so buys you a 3Ware 9650SE PCIe SATA RAID controller and the optional battery backup unit.

-Brad

I keep seeing people suggest RAID 0+1, when what you actually are interested in is RAID 10. They are not the same thing, and RAID 0+1 is inferior to RAID 10.

Of course, RAID 5 is superior to them both, provided that the controller card cost is not an issue.

Thanks guys! I'm going to ghost the drive and set it away as it will only be apps. All data, songs, samples ect will be backed up by band, artist and song nightly so thats not a problem.


Woudl RAID 0 be the easiest way to make 2 drives act as one? I know this won't give me any safety net, but I will be backing up to external drives, and eventually Blueray discs.


Thanks so far for all the support! 


Blake

I keep seeing people suggest RAID 0+1, when what you actually are interested in is RAID 10. They are not the same thing, and RAID 0+1 is inferior to RAID 10.

Of course, RAID 5 is superior to them both, provided that the controller card cost is not an issue.
Please don't make assumptions about what you think I meant.

There is absolutely, positively nothing about RAID 5 that is superior to RAID 0+1 or RAID 10 unless cost is your primary issue. The resiliency of a RAID 5 set would be no better (and in fact would be worse than RAID 10) and the performance can't possibly approach RAID 0, RAID 1, RAID 0+1 or RAID 10.

As for RAID 0+1 vs RAID 10, the only thing superior about RAID 10 is an extra degree of resiliency. However with the proper controller and driver a RAID 0+1 can enhance performance even further by permitting simultaneous reads and writes. In that case - the difference between logically mirroring a stripe and striping a mirror, there's a small tradeoff between performance and that extra degree of resiliency.

You may indeed be correct however, as I don't know for sure that the 3Ware supports that sort of multithreaded access. For $300, perhaps not...

-Brad

Raid 0 (which really isn't "raid" but that's irrelevent to the conversation), Raid 1, and Raid 0+1 are all supported by modern motherboards. Raid 10 is only supported by a select few.

RAID 1 will actually virtualy merge the 2 drives though right?

Raid 1 is Mirroring, which basically looks to the system as one hard drive, but creates two exact copies of the info on the two drives. I think you'd be best off with the 0+1 or 10 options listed, depending on what you're looking for. "Raid" 0 is Striping, which treats two hard drives as one hard drive, however, if either fails, all your info is gone, which is why people use the 0+1 and 10 options whch have both the above features although the process the info goes through is slightly different, but that's not something an enduser should be too worried about. Use 0+1 if you can't afford a way to use 10. Raid 5 will be slower than the other two and you can use 6 if data is absolutely crucial (which doesn't seem to be the case in this instance)

I didn't assume anything about what you meant. I based my response on exactly what you wrote. You suggested RAID 0+1 for both speed and redundancy, and without having to spend an arm and a leg on a controller. And I disagree with that recommendation -- those requirements are met with greater resiliency using RAID 10.

RAID 0+1 offers no performance advantage over RAID 10. Neither needs beefier controller requirements than the other. And RAID 10 is more resilient in the various failure modes. The advantage is RAID 10 for that specific scenario.

RAID 5's advantage over both is a more efficient use of the available raw hard drive space. Both RAID 0+1 and RAID 10 lose half the hard drive space, which is a significant loss of money if using Raptors. Less hard drives = less heat, less power, and less chance of individual hard drive failure. RAID 5's read speed using a good controller is comparable to n-1 drives in a RAID 0, faster than RAID 1, and only slightly slower than RAID 0+1 or 10. Write speed, of course is where you may take the performance hit, and this is dependent on the controller. The 3Ware 9650SE can sustain over 100MB/sec RAID 5 writes. Resiliency of the RAID5 array is good, equal to the resiliancy of RAID 0+1, but RAID 10 can handle some multiple hard drive failure modes that RAID 5 can't. Of course, the 3Ware 9650SE card can now do RAID 6 which can handle 2 hard drive failures, at only a 20% write speed penalty over RAID 5.

The benefit/trade-off of RAID 5 is that you have to spend more money on the controller, but potentially less money on hard drives for the same space, speed, and safety. I like RAID 5's simplicity of setup, and the reduced drive count over RAID 0+1 & 10 translates into a heat & power advantage.

I will say that IOPs (I/O's per second) on RAID 0+1/10 probably perform much better than a typical RAID 5 implementation. This translates into better performance for high-transaction applications like databases. However, for System6's audio applications, it would be dubious if this is an advantage. In fact, with audio applications (working with large files like .wavs), transfer rate would be a better benchmark of his expected performance gain than anything else.

Perhaps Seagate 7200.10 drives (perpendicular recording, 7200 RPM) in a RAID 0, RAID 10, RAID 5, or RAID 6 are a good solution for his system. Depends on how much redundancy he wants and how much money he wants to spend on a controller.

You were the one who mentioned controller cost, not me.

The rest of your post looks fairly good.

By the way, since the OP claims to be running a recording studio, I would be inclined to believe that write speed would be a significant matter in multitrack recording. Worse yet, if a drive in a RAID 5 set fails during production not only is I/O slowed due to the rebuild process once the spare is put online, but all the I/O is slowed even further due to parity operations. At least in a mirror set the only slowdown will be the rebuild process and that can be made low priority. All the more reason to shy away from RAID 5 in my opinion.

-Brad