Noob question - 3 gb/s in SATA description?

dbaldus

Gawd
Joined
Jul 12, 2005
Messages
563
So I always see "SATA 3Gb/s" in the description of drives, but every time I hear someone talking about their new SATA II drives, they always talk about peak bursts at like 160 mb/s or something like that.

I'm just confused as to what the 3 gb/s actually means. I assumed that was the transfer rate, but then why do people talk about 160 mb/s bursts? Also, are all of the original SATA drives 1.5 gb/s? What are the rates on old IDE drives, such as my Maxtor Diamondmax 9 ATA 133?

I'm just trying to decide if it is worth it to step up to a SATA II model or not so I would like to see what these numbers really mean.
 
In both caes, 3gbps and ata 133 are in reference to the maximum theoretical bandwidth of the interface, no drive can hit that with a sustained rate on either platform, though some scsi drives are getting much closer to those throughput rates.
 
So what do SATA II (3 gbps) drives typically do?

And what do ATA 133 drives typically do?
 
The ATA133 is 133 megabytes per second (maximum theoretical sped of PCI 33 bus so never achievable even if the disk could do it)

are the SATA interface advertising "Gb/sec is giga bits so divide it by 8 and to get GB per sec. (small b means bits big one means bytes) then slow it down a bit more for conversion.

anyways by the time you do all the math the "3Gb/sec" would be about the same as ATA300 if it existed.

still waiting for a drive that can sustain ATA100 :)
 
DeFex said:
The ATA133 is 133 megabytes per second (maximum theoretical sped of PCI 33 bus so never achievable even if the disk could do it)

are the SATA interface advertising "Gb/sec is giga bits so divide it by 8 and to get GB per sec. (small b means bits big one means bytes) then slow it down a bit more for conversion.

anyways by the time you do all the math the "3Gb/sec" would be about the same as ATA300 if it existed.

still waiting for a drive that can sustain ATA100 :)

Whoa... wait a second...

So the SATA II drives can't even sustain 100 megabytes per second? Which would mean that SATA I drives can't sustain 50 megabytes per second? And where does that put IDE drives??

And back to my original question... what DO IDE drives typically perform at anyway? Trying to figure out my increase in transfer speed, on average, between:

1. An IDE drive and a Sata I drive
2. An IDE drive and a SATA II drive

And then decide if it is worth the upgrade.

Thanks for the info so far!
 
You misunderstanding.

The fastest sata drive right now is still teh 10k raptors, they can sustain about between 90-65 MB/s depending on where you're at on the platter. Most 7200 rpm drives handle around 70-45 MB/s these days.

If you take a drive on SATA II (theoretical max 300 MB/s) that can sustain 70-45 MB/s it will also sustain 70-45 MB/s on SATA I (theoretical max 150MB/s).

No current drive exsists that can saturate the speed of it's intended interface except possibly ram drives. Also, there are no drives, except ramdrives, that can acheive rates of 150MB/s sustained.

If you take three identicaly drives except for the interface, IDE and SATA I, and SATA II they will perform with in a few percentage points of each in rates. The only thing affected by the interface in that situation would be the latency. We're talking 1-2 % difference though, something not even noticable during average, everyday use.
 
Ah... starting to become clear now. It is really the drives that are the inhibitors for speed, not the interface, because the drives perform nowhere near what the interface would allow. At least thats what I interpret from the above post.

So what I should really be asking is...

How fast is my Maxtor DiamondMax9 ATA 133 performing, and how fast would Seagate Barracuda 7200.10 3 Gb/s perform, on average (as opposed to how do IDE drives perform against SATA II drives). Does that sound correct?

Anyone have any experience with the speed of DiamondMax9 drives? How bout the Seagate Barracudas? Thanks!
 
dbaldus said:
Ah... starting to become clear now. It is really the drives that are the inhibitors for speed, not the interface, because the drives perform nowhere near what the interface would allow. At least thats what I interpret from the above post.
Performance isn't measured in megabytes per second. They're a determining factor, but only in highly sequential applications do you see a difference in performance proportional to STR.

Second, the same drive has the same capabilities, regardless of interface. Think of the interface as a limiting speedometer on a car - you could put a speedometer that goes up to 300 MPH on a slow car, and it wouldn't improve it any. But if you took a really fast car and put on it a 33 MPH speedometer, you wouldn't be able to go faster than 33 MPH because the speedometer would slow it down. Since no drive is really hitting 150 MB/s, "sata 2" speeds don't help anything when you've got the standard point-to-point topology of sata. With port multipliers, it'd be a concern - five disks can certainly STR more than 150 MB/s - but not many people are using them yet that I've seen.

Check storagereview.com for a comparison of application-level performance of many drives and a good number of applications.
 
unhappy_mage said:
Performance isn't measured in megabytes per second. They're a determining factor, but only in highly sequential applications do you see a difference in performance proportional to STR.

Second, the same drive has the same capabilities, regardless of interface. Think of the interface as a limiting speedometer on a car - you could put a speedometer that goes up to 300 MPH on a slow car, and it wouldn't improve it any. But if you took a really fast car and put on it a 33 MPH speedometer, you wouldn't be able to go faster than 33 MPH because the speedometer would slow it down. Since no drive is really hitting 150 MB/s, "sata 2" speeds don't help anything when you've got the standard point-to-point topology of sata. With port multipliers, it'd be a concern - five disks can certainly STR more than 150 MB/s - but not many people are using them yet that I've seen.

Check storagereview.com for a comparison of application-level performance of many drives and a good number of applications.

Nice. Thanks!
 
Back
Top