- May 19, 2011
- 19,108
- 12,361
- 136
Yes, it's true that there aren't any normal HDDs that can even peak anywhere near what SATA 3.0 (6Gbps) is capable of, but I would like to see a review of today's SATA 3.0 HDDs benchmarked on a decent SATA 3.0 controller, then on a decent SATA 2.0 controller, then in SATA 1.0.
I think what people are missing when they make statements along the lines of the title of this thread is that HDDs have never challenged the top speed of the bus they were designed for.
Example. When UDMA33 was the current standard, there weren't any IDE disks that could make full use of the available bandwidth (Peak or sustained, though personally I feel that sustained bandwidth is the only figure that really counts for anything practically).
UDMA100 must be at least a decade old now, I have a computer with UDMA100 and the board must have be made around 1999, yet there's another computer sitting here, Core i7-920 with two SATA 2.0 capable disks, transferring a large file between them and it's going at... 100MB/sec. However, if you found a relatively decent-spec computer with two UDMA100 hard disks in, I estimate that one would be fortunate to have them managing a sustained transfer speed of say 10MB/sec.
In my experience it took until about 2006 before I started seeing sustained transfer speeds of 50MB/sec between two internal disks, yet even SATA 1.0 is only capable of around 150MB/sec.
Of course, a number of things affect the speed at which a disk can performed sustained transfer speeds. Most of them are to do with how much data is packed onto each platter and other physical attributes, however without the advances of the ATA specs, improved efficiencies in signalling between drive and controller, etc, a lot of what a drive could do would mean very little.
I think what people are missing when they make statements along the lines of the title of this thread is that HDDs have never challenged the top speed of the bus they were designed for.
Example. When UDMA33 was the current standard, there weren't any IDE disks that could make full use of the available bandwidth (Peak or sustained, though personally I feel that sustained bandwidth is the only figure that really counts for anything practically).
UDMA100 must be at least a decade old now, I have a computer with UDMA100 and the board must have be made around 1999, yet there's another computer sitting here, Core i7-920 with two SATA 2.0 capable disks, transferring a large file between them and it's going at... 100MB/sec. However, if you found a relatively decent-spec computer with two UDMA100 hard disks in, I estimate that one would be fortunate to have them managing a sustained transfer speed of say 10MB/sec.
In my experience it took until about 2006 before I started seeing sustained transfer speeds of 50MB/sec between two internal disks, yet even SATA 1.0 is only capable of around 150MB/sec.
Of course, a number of things affect the speed at which a disk can performed sustained transfer speeds. Most of them are to do with how much data is packed onto each platter and other physical attributes, however without the advances of the ATA specs, improved efficiencies in signalling between drive and controller, etc, a lot of what a drive could do would mean very little.
Last edited:
Facebook
Twitter