Sonar Producer 8: Separate hard disk for audio data needed?

I have always kept my audio drive separate from my program drive as a practice.
 
Need? It depends on what you are doing with your computer and how fast it is.

Would it also be more efficient if I partitioned the hard drive to create a separate partition just for the audio files?

Nope. Here's the deal: the slowest activity that a disk drive performs is called a seek, meaning the action of physically positioning the head over the desired track in preparation for reading and writing data.

Compared to reading and writing data, seeking happens on a glacial timeframe. Minimizing seek times is therefore the single most important technique for speeding up disk drives. That's why you defrag a disk, so that pieces of large files are physically near one another, thereby reducing seek times.

Once the head makes it to the track, data can be sucked off the disk quite rapidly. But once the track's been read (or written), then the head needs to seek again to the next file fragment and everything slows to a crawl again.

Because drives have multiple heads, it can read more than one track before it has to move the heads again. So when data is written, all the heads are used sequentially. A group of tracks that can be read or written from the same head position is called a cylinder (from back when you had one head per platter surface, so that a vertical set of tracks formed a cylinder). The larger the cylinder, the more data that can be moved between seeks.

When you partition a drive, you're making smaller cylinders. Partitioning makes sense only for higher-level data organization and not for efficiency, which actually suffers. For maximum speed, you should always make the entire drive one partition.

Whenever a program must access multiple files concurrently, it has to do a lot of seeking as it furiously bounces from one file's physical location to another's. But if those files are separated onto different drives, each having its own set of read-write heads, one drive can be seeking while another is reading or writing data. Fewer seeks, faster performance.

Bottom line: use multiple physical drives, each formatted into a single partition.

Thanks Bitflipper!  That was interesting...and thanks for the detail.  You explained some stuff I had wondered about before.

they say internal SATA is the way to go, so that's the way i went, and it is all good.

Bitflipper,
 
I've been trying to learn more about your statement in a post on separate audio drives from 2009.
You said: "When you partition a drive, you're making smaller cylinders. Partitioning makes sense only for higher-level data organization and not for efficiency, which actually suffers. For maximum speed, you should always make the entire drive one partition."
 
I've been trying to research this, but I can't find anywhere that "partitioning a drive makes smaller cylinders".
 
If I understand correctly, a cylinder is made up of  tracks that are lined up vertically underneath each other on the platters. Partitions are created in horizontal concentric circles on all platters, moving from the outside to the inside of the platters. So how would making a partition reduce the size of the cylinders? As far as I can see, it would only reduce the number of cylinders, not the size.
 
I understand that it's not good to have frequently accessed data on two separate partitions, because this would cause the seek heads to shuttle back and forth between the partitions, thereby slowing performance. But what about if you had frequently addressed data on the outside partition, like audio data, but the other, inside partition you used only for storage and backups? You would never be using both partitions at the same time, so I can't see how this would degrade performance . . . unless you are right about cylinder size being reduced.
 
Can you please clarify?
 
Thanks
 
 

The term "cylinder" is a holdover from the old days when you had one head for each surface. Back then, the logical organization really could be visualized as cylindrical. Partitions were horizontal slices of cylinders. Nowadays, of course, you have multiple heads for each platter, so there isn't anything truly cylindrical about it anymore.
 
The underlying concept, however, is still valid. Although the multiple read/write heads can function independently, they are all still attached to a common arm and move together, even though they may be positioned over disparate logical partitions.
 
When reading data, each head reads a portion of the "cylinder" in sequence. Then the heads are moved to the next physical location. Seeking is the most expensive operation a drive can perform, so the fewer the number of seeks the faster you get the data. The larger the partition, the more heads are involved in reading data before the next seek, and the more data you can move before physically re-positioning the head assembly.
 
Hope that makes sense. Believe it or not, I once taught the subject to field engineers! But it's been a long time since the days when I sat on a computer room floor with a dismantled washing-machine sized 300MB CDC drive.