On Tue, 19 Dec 2017 22:54:19 -0600, VanguardLH wrote:

Mark Perkins wrote:

On Tue, 19 Dec 2017 20:35:33 -0600, VanguardLH wrote:

Percival P. Cassidy wrote:

http://www.tomshardware.com/news/hdd...ate,36132.html

Defragmenters will have to adopt new algorithms so they position
sequential data not in successive clusters within the same track on the
same platter but jigsawing the data across multiple platters. So one
head can read one cluster of data while a head on the other arm can be
reading the next logically sequential cluster of data in building up the
data stream.

I didn't get that impression from the article.

From the blog post:
The host computer can treat a single Dual Actuator drive as if it were
two separate drives. This means the host computer can ask a single
high-capacity drive to retrieve two different data requests
simultaneously — delivering data up to twice as fast compared with a
single-actuator drive.
/blog

The host computer is expected to ask the drive to do two things at once,
but I don't see that as expecting the drive to read interleaved data and
assemble it on the fly. Instead, it sounds more like the following:

read two files simultaneously
write two files simultaneously
write one file while reading another

Besides, I'm not worried about defragging. That went out of fashion many
years ago.

With your interpretation, there would be nothing faster about reading or
writing *a* file. Not much point in wasting the money on this drive
technology.

Statistically, when is a modern computer reading or writing a single
file? Almost never, I'd say. So, yes, I think this will be significantly
faster, at least in theory. Who knows, it may move the bottleneck
somewhere else.