SATA not much faster than PATA yet?

[wizzzer]

OK like SATA is supposed to be capable of 150mb/sec right?
And SATA II 300mb/sec, right?

Well I just looked at a table of hard drives in October 2005 PC World,
and it shows that in real world testing, the fastest PATA drive could
copy a file at about 15mb/sec and the fastest SATA drives were
something like only a third faster than that.

So right now it seems like this SATA hype is just bogus, doesn't it?
It seems like they just made new standards to get people to buy new
hardware.
Seems like since computer hardware is so long-lasting, they keep
releasing lots of new standards to get people to keep upgrading.

 

[VWWall]

OK like SATA is supposed to be capable of 150mb/sec right?
And SATA II 300mb/sec, right?

Well I just looked at a table of hard drives in October 2005 PC World,
and it shows that in real world testing, the fastest PATA drive could
copy a file at about 15mb/sec and the fastest SATA drives were
something like only a third faster than that.

For years the electronics was the limiting factor in hard drive speed.
As the drive's recording methods got better, new specs like
ATA66/100/133 were devised. Above ATA33, 80 wire cables were required,
and even with these. The ATA133 spec was about the best parallel drive
electronics could support. As you mention, few PATA drives can use the
full ATA133 speed.

Any drive performance depends on how fast the "bits" can go by the
head(s). For present drives, this means RPM is the best indication of
speed. New methods of magnetic recording will allow for at least a 10
fold factor in density, and hence speed.

So right now it seems like this SATA hype is just bogus, doesn't it?
It seems like they just made new standards to get people to buy new
hardware.

There are other advantages to SATA drives, but until the drive mechanism
is upgraded, the full speed advantage of even SATA I will not be seen.

Seems like since computer hardware is so long-lasting, they keep
releasing lots of new standards to get people to keep upgrading.

I think the software is worse than the hardware!

 

[kony]

OK like SATA is supposed to be capable of 150mb/sec right?
And SATA II 300mb/sec, right?

No.
That's only a theoretical bus transfer rate before
considering any overheads... similarly any bus in a computer
is capable of more than actually delivered, but in this case
SATA is particularly bottlenecked by the mechanical hard
drive itself.

Well I just looked at a table of hard drives in October 2005 PC World,
and it shows that in real world testing, the fastest PATA drive could
copy a file at about 15mb/sec and the fastest SATA drives were
something like only a third faster than that.

Then their test was poorly done, usually the difference
between SATA and PATA is much lower, but both transfer well
over 50MB/s with a fast/modern drive. Just goes to show
that a test/review/etc/etc is only as good as the person
writing it.

So right now it seems like this SATA hype is just bogus, doesn't it?
It seems like they just made new standards to get people to buy new
hardware.

It's evolutionary... like anything else, baby-steps forward,
not giant leaps.

Seems like since computer hardware is so long-lasting, they keep
releasing lots of new standards to get people to keep upgrading.

If it's so long-lasting, why would you keep upgrading?
Nobody's twisting your arm.

Fact is, eventually PATA bus would be a bottleneck so it's
better to increase the bus rate as much as possible now
rather than waiting till it is. Then there were the smaller
cables, longer cable lengths... Serial connections are going
to be the preferred interface when possible.

 

[Paul]

OK like SATA is supposed to be capable of 150mb/sec right?
And SATA II 300mb/sec, right?

Well I just looked at a table of hard drives in October 2005 PC World,
and it shows that in real world testing, the fastest PATA drive could
copy a file at about 15mb/sec and the fastest SATA drives were
something like only a third faster than that.

So right now it seems like this SATA hype is just bogus, doesn't it?
It seems like they just made new standards to get people to buy new
hardware.
Seems like since computer hardware is so long-lasting, they keep
releasing lots of new standards to get people to keep upgrading.

You can get some real data here:

http://www.storagereview.com/comparison.html

Select "Maximum Transfer Rate" to see the disk speed at the
beginning of the platters. Select "Minimum Transfer Rate" to
see the disk speed at the end of the platters. Compare those
results to your "PC World" article.

Paul

 

[Ian East]

OK like SATA is supposed to be capable of 150mb/sec right?
And SATA II 300mb/sec, right?

Well I just looked at a table of hard drives in October 2005 PC World,
and it shows that in real world testing, the fastest PATA drive could
copy a file at about 15mb/sec and the fastest SATA drives were
something like only a third faster than that.

So right now it seems like this SATA hype is just bogus, doesn't it?
It seems like they just made new standards to get people to buy new
hardware.
Seems like since computer hardware is so long-lasting, they keep
releasing lots of new standards to get people to keep upgrading.

The theoretical speed of the bus is more or less meaningless since
it's many times greater than the mechanical limitations of the disk
itself. The only real thing that matters for performance is the size
of the cache and rotational speed of the hard disk. A 15K rpm SCSI
disk with a 16MB cache will easily beat the pants off a 8MB cache 7200
rpm SATA disk any day even if it's a 80MB/s SCSI bus.

However SATA is supposed to support tagged command queing (something
SCSI and FibreChannel disks have been doing for years) but in practice
most SATA controllers and disks havn't implemented it yet. This
basically allows the disk to reorder disk read and write requests so
that the drive head requires minimal movement.