Sata 1 and Sata 2

M

MS

Hello, I have a question about the specs for sata.
1) Is sata 1 max. speed of 1.5 Gbits per second or 1.5 Gbytes per second?
The thing is that I have a Raid 0 with two Samsung HD. This system reports
in Sandra XL a drive index of 186 MBytes per second. Now, if we divide
1500/8 (Sata 1 1500MBits/s / 8 bits) the result is 187.5 MBytes/sec. This
means my system is Sata 2 compliant but is running in Sata 1 mode.

MS
 
P

Paul

MS said:
Hello, I have a question about the specs for sata.
1) Is sata 1 max. speed of 1.5 Gbits per second or 1.5 Gbytes per second?
The thing is that I have a Raid 0 with two Samsung HD. This system reports
in Sandra XL a drive index of 186 MBytes per second. Now, if we divide
1500/8 (Sata 1 1500MBits/s / 8 bits) the result is 187.5 MBytes/sec. This
means my system is Sata 2 compliant but is running in Sata 1 mode.

MS

Sata I = 1500 Megabits/sec on the SATA cable. Coding is 8B10B, meaning an
80% efficient code is used. Send 1500, get 1500 * 0.80 = 1200 Megabits/sec
at the user level. Divide that number by 8, gives 150MB/sec at the
user level.

The protocol is not 100% efficient, so the cable cannot deliver at exactly
150MB/sec. There will be gaps, periods of inactivity. The "burst transfer
rate" provided by some test utilities, will give some idea how efficient
the actual protocol is. Like many high speed protocols, packets are involved,
so there is some information in the header of the packet, which "wastes"
bandwidth. You would not expect to see the number "150" show up in
a benchmark, it'll be a slightly lower number taking into account
the losses due to packet headers.

I was actually trying to find a HDTune benchmark a few hours ago, to see
if I could get a burst transfer result for SATA I, but I couldn't find
an example. For such a measurement to be useful, the hardware could
not have any other bottlenecks to get in the way. For example,
it would be silly to use a PCI SIL3112 card to do such a test,
because the PCI bus bandwidth is a limit in that case. I had to reject
some of the graphs, because of the possibility that the hardware
was a limitation.

Your RAID0 benchmark, if you were measuring sustained transfer rate,
might be showing the head to platter limit. Take 186MB/sec and
divide by two, and that would be 93MB/sec coming from one of the
Samsung disks. You could try the Storagereview performance database,
to compare the result against what they've measured. They have quite
a few drives, but maybe they haven't checked your Samsung.

http://www.storagereview.com//Testbed4Compare.sr (site uses popup adverts)

Paul
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top