SATA, 3G Questions
I'm using a SATA3G DFI Mobo and have some questions about using SATA3G and I devices together.
If a SATAI HD is primary, and a SATA3G is secondary; will the SATA3G HD only run at SATAI speeds?
If a file transfer from SATA3Gs to SATAI drive is initiated, would the read from SATA3G only run at SATAI speeds?
Specific HD's I'm looking at are the SATAI WD Raptor 73GB & SATA3G Seagate 500gb
EDIT: I fixed the typos and confusion about SATAII/SATA3G
If a SATAI HD is primary, and a SATA3G is secondary; will the SATA3G HD only run at SATAI speeds?
If a file transfer from SATA3Gs to SATAI drive is initiated, would the read from SATA3G only run at SATAI speeds?
Specific HD's I'm looking at are the SATAI WD Raptor 73GB & SATA3G Seagate 500gb
EDIT: I fixed the typos and confusion about SATAII/SATA3G
0
Comments
The advantage with the SATA3G is that the entire SATA controllers throughput is 3Gb/s.
I can see 3 HD's transfering information at the same time can easily overwhelm a SATAI controller; going over 150mbs.
The question is if there are SATAI devices in conjunction with SATA3 devices, would the controller throttle down to SATAI? Or would the SATAI interface on that drive simply limit it to the 150mbs and the total bandwith on the SATA controller(including all other HD's bandwith) still be 3Gbs?
No, each channel has 3Gb/s. Each channel of a SATA1 controller has 1.5Gb/s. the quoted bandwidth refers to each channel, not the total bandwidth of the controller.
Your right there unless your goina have 10 drives in raid 0 you will never max out satas bandwith.
yes, but before editing the post was asking about SataII and sata3 as if they were 2 seperate standards.
However, this discussion does raise the question for me- I've seen people say that overall SATA controller speed is still restricted by the slowest drive plugged in. But with the removal of the dependencies generated by having multiple drives on the same cable and the same single clock (they now use the data stream embedded clock method), I thought that this no longer has to be an issue?
Based on that- I would think that if your read data from the 3gbs it would happen at that speed (if the controller supports it) and would write it at 1.5 to the 1.5 drive. Though the overall transfer rate would STILL be 1.5 (best case)- each drive could(?) handle it at their indiviual rate? (EDIT ADDED: The standard does say that SATA uses a a point-to-point topology and has the capability to work independently and NO sharing of interface bandwidth. But how that is working in practice ...? :END ADD)
Methinks that it is a function of how your controller is built.
At any rate, it's turned out to be a good question I've wondered about-
Ignoring that though, if the controller is reading data from one drive at 3Gb/s and writing to the other at 1.5Gb/s where is the rest of the data going? If you're writing to your destination drive at 1.5Gb/s it stands to reason that the fastest you can read from your source is 1.5Gb/s. It's not a function of the controller or any inherent limitation in SATA, it just stands to reason.
So does the controller throttle down to SATAI just because one of the HD's are at SATAI speed, or is that specific I/O on the HD going to limit just that HD?
Also, the "confusion" is simply the answer to the question if a 1.5gbs drive does limit a 3gbs (ADDED: on the same 3gbs capable controller) . The answer according to the standard is NO, not necessarily. The answer in practice, I do not know. I thought I've read posts where people thought it did. That's where I'm fishing (both here and on the web).
Unfortunately, you won't get that many drives off any single controller on a DFI board- and hence deicist's points are valid. Each controller has a limit of 4. Even if one could, I'm uncertain.
However, you might get that many on with a port multiplier like the SiI 3726 which DEFINITELY supports the asynchronous timing you are asking about, but I think it is a solution that currently only works with the 3124-2 PCI-X solution.
Others may have an idea, but that is all I could find at this time.
This might have interesting convergence implications too, D.
Remember with PATA Ultra100 controllers that each channel was only as fast as the slowest drive connected to that channel but it only affected that channel! Well with SATA and SATA 3.0 controllers each channel only has one drive attached. So this means (and I have personally tested it in many combinations) that each channel is going to run independent of the of the other channels. It does not matter what combination you run as long as you are not running different drives in RAID. I currently have 2 74G Raptors (not in RAID) and 2 Hitachi SATA 3.0's (in RAID-0) all connected to the same controller on my DFI nF4 SLI DR and all is well under the hood!