Sign in with
Sign up | Sign in
Your question

SATA II does not exist. Ignore if you already know this.

Last response: in Storage
Share
April 30, 2006 8:29:14 AM

If you already agree with the Subject line just ingrore the

SATA II was the name of the standards group which they sense changed to SATA-IO.

They specifically demand than no one use the term SATA II for drive supporting 3.0 Gbps rates. SATA II means nothingl. If anything it would apply to all exisitng SATA drives.

--

But isn't SATA 150 MBps equivalent to generation 1 in effect SATA I
while SATA 3.0 Gbps equivalent to generaltion 2 in effect SATA II

---

A common myth is that SATA 3.0 Gbps supporting drives are the next gneration and have advanced features absenst from the earlier SATA 150 MBps.

Nothing could be further from the truth.

1) No SATA drive can break the 100 MBps barrier so increasing the interface's maximum limit to 300 MBps has no practical value!

2) Supporting the 3.0 Gbps transfer rates in no way indicates what other feature are choses. Some SATA 150's feature NCQ while some SATA 3.0's don't.

So drives cannot be divided into neatly into two classes based on maximum interface speed.

3) Western Digital has chosen to use SATA 150 for all their top of the line products like the new 150 GB Raptor. So SATA 3.0 does not be newer or better.

4) Some people claim 3.0 Gbps drives are faster, that maybe the interface is more efficent, or that they have caches that exceed 150 MBps giving them a speed boost. I have seen no evidnece of this and the fact that WD choses not to updgrad the 150 Raptor with a 3.0 Gbps interface while including the same interface on their non-Enterprise lines suggest to me that there is not benefit- at least not until hard drives get about 50% faster.

As far as the transfer rates go its just like ATA 100 vs ATA 133.

When evaluating hard drive's just ignore the SATA interface speed and look at the benchmarks and the supported features.

http://www.storagereview.com/php/benchmark/bench_sort.php

BTW if you think I am full of crap check out the SATA-IO's homepage. You can't get much more authoritative than that.
http://www.sata-io.org/namingguidelines.asp

More about : sata exist ignore

April 30, 2006 12:11:57 PM

GOD dam... I should have just got a SATA I HD instead of a SATA II HD! Although there was only a $10 AUS difference between the two HDs. However I think I got NCQ for my additional 10 bucks.
April 30, 2006 6:23:51 PM

Quote:
GOD dam... I should have just got a SATA I HD instead of a SATA II HD! Although there was only a $10 AUS difference between the two HDs. However I think I got NCQ for my additional 10 bucks.


And NCQ really don't help in desktop environnement...
Related resources
April 30, 2006 8:45:36 PM

Quote:

And NCQ really don't help in desktop environnement...


Not everyone agrees, with others saying single-user benefits are large, while less than multi-user. Check out this review:

http://www.sharkyextreme.com/hardware/article.php/35279...

For what it's worth, Seagate's official marketing claims single-user benefit too:

"The Barracuda 7200.8 with NCQ benefits applications like high-performance PCs such as gaming systems and workstations, PC-based home media servers and entry-level SATA servers. In addition, NCQ can make booting, application loading and file copying faster for everyday mainstream users."

It also claims less wear on the drive.
April 30, 2006 9:28:07 PM

Quote:

1) No SATA drive can break the 100 MBps barrier so increasing the interface's maximum limit to 300 MBps has no practical value!


So you don't think they should come out with standards for next gen stuff before the next gen stuff arrives? Supposedly perpendicular hard drives will break the 100MB/s mark soon, and should go up from there. I'm guessing when they start using DDR2/3 for the ram drives those should easily hit even 300MB/s.

And isn't SATA II supposed to be both ncq and 300MB/s? I thought it was in the spec, although you can add stuff to lower standards if something says sata II is better have ncq.
April 30, 2006 9:52:37 PM

SATA specifications are theoretical bandwidth. SATA I is 150 and SATAII is 300.
"1) No SATA drive can break the 100 MBps barrier so increasing the interface's maximum limit to 300 MBps has no practical value!"
Why you use RAID. 2 drive RAID and break this barrier on AVERAGE read and BURST is higher. Three and four drives AVERAGE read is much higher while BURST goes up at an incredible rate. Looking at these charts, BURST is sustained through most of the drives platters. Just like a CD, slow at first then it picks up speed as the head moves outward.
"2) Supporting the 3.0 Gbps transfer rates in no way indicates what other feature are choses. Some SATA 150's feature NCQ while some SATA 3.0's don't. "
Some benefits are gained from NCQ. Is it really noticable? Good question. From my own test, I can't find where it is noticible, but I can't test everything.
2) Supporting the 3.0 Gbps transfer rates in no way indicates what other feature are choses. Some SATA 150's feature NCQ while some SATA 3.0's don't.
Turning my HDs back to SATA 150 resulted in loss of speed, not a terrible amout but noticable, especially on large files.
In some examples I agree, while in others I don't. Comes down to this: Do you really need it? For me, and my applications, yes. Just like the move from AGP 4X, 8X and now PCIe, you have the potential for greater headroom. Can we use it all? No, but it's there when or if the technology is available. Another reason to use RAID.
April 30, 2006 11:12:22 PM

Quote:
So you don't think they should come out with standards for next gen stuff before the next gen stuff arrives?


Nope I am all for it. Didn't mind a bit when ATA 133 was introduced, until people started assuming ATA 133 drives where better or advanced than ATA 100 and stating their assumptions as fact.

Quote:


And isn't SATA II supposed to be both ncq and 300MB/s? I thought it was in the spec, although you can add stuff to lower standards if something says sata II is better have ncq.


If that were true then I would be a complet idiot who is just arguing semantics.

Let me make this very clear.

Not only does SATA I and SATA II products don't exist there is nothing out there that is functionally equivalent to there beging two levels of SATA!

All optional features defined by SATA standards are seperate.

There is no requirment or even a recommentation that when you advertise feature X you should also implement Y or Z.

It neither hurts nor helps for a drive to support the 3.0 Gbps tranfer rate.

Having the 3.0 Gbps tranfer rate doesn't make it more or less likely to have other advanced SATA features that may actually matter.

--

Some manufacturers are using deceptive practices to convince people to buy their product. A lot of hobbyists assume when they see two labels being applied to the product, one faster than the other that there are two levels or two generations availiable one better than the other.

Then a few of the hobbyists start presenting their guesses as facts without even considering taking 5 minutes to check out the SATA-IO website. Many people hear the same wrong infromation from many different sources and they start to belive its accepted fact.
April 30, 2006 11:53:26 PM

Quote:
SATA specifications are theoretical bandwidth. SATA I is 150 and SATAII is 300.
"1) No SATA drive can break the 100 MBps barrier so increasing the interface's maximum limit to 300 MBps has no practical value!"
Why you use RAID. 2 drive RAID and break this barrier on AVERAGE read and BURST is higher. Three and four drives AVERAGE read is much higher while BURST goes up at an incredible rate. Looking at these charts, BURST is sustained through most of the drives platters. Just like a CD, slow at first then it picks up speed as the head moves outward.


I am trying not to be rude but

You either don't understand how RAID works or you don't understand how Hard Drives work.

1) RAID doesn't magically make the indidual drives transfer data faster, it works by splitting the load between individual drives.

2) The transfer rate is for the link between each individual drive and the controller. For PATA that link may be shared by a master and a slave. For SATA one drive gets the whole 150Mbps-300Mbps.

3) AGAIN no SATA drive breaks the 100 MBps barrier! And yes I know the difference between maxium transfer rates and average rates. There is currently a 15000 RPM SCSI drive that peaks slightly above the 100 MBps barrier. Have you ever looked at any hard drive benchmarks?

---
And before someone else misreads a benchmark and brings this up.

Hard drives have a 2-32 MB of RAM on them used as a cache. This cache can be read much faster than transfering data from the drive proper.

However.

1) The hard drive can't move data to 2-32 MB or RAM any faster than it can move data over the bus. Its a bus stop not an additional lane on a highway.

2) The OS does its own caching. So if you pull 2 MB from the HD and then need the same 2 MB from it again, its still going to be in your system memory so it doesn't matter that you can pull it from the HD's cache rather than from the HD.

The cache does have an important function, but its not speeding up transfer rates.

---
Let me reitterate.

1) There are currently no SATA compliance levels. Its a grab bag of features and having one says nothing about having another.

2) Just becuase a drives says SATA 3.0 Gbps doesn't automatically make it faster, better or more advanced than a drive which says SATA 150. So for now you can make no assumptions, each drive has to be evaluated on its own merits.

3) Its IS better for a controller to support SATA 3.0 Gbps, becaues there are already existing products which could easilylbreak 150 MBps and more are likely on the way.

Gigabytes i-RAM would definately benifit from switching to SATA 3.0 Gbps and I would be supprised if 3.0 Gbps version wasn't already in the works.

You could also make a RAID controller which connected to a SATA port rather than the PCI express. This would have the benifit of appearing to the system a just a single fast drive without needing special driver support.

The same can be said of external enclosures employing RAID that using the SATA interface.

4) Someday soon I expect drives that actually need SATA 3.0 Gbps. Yes going from the 80's to above 150 will likely take a long time, but there are also plans to make hybrid drives that include NVRAM.

5) Probably by the time 3.0 Gbps is useful to a hard drive there will be 6.0 Gbps drives being marketed. (its already part of the standard).
May 1, 2006 12:27:35 AM

I'm trying not to be rude either but you make it hard not to. I know how RAID works as I've used them for many years. Obviously, you do not. You can't seem to see past a benchmark that shows a single HD with less than 100MB read.

1) "RAID doesn't magically make the indidual drives transfer data faster, it works by splitting the load between individual drives. " Correct, creating faster reads/writes, depending on the array.

3)"AGAIN no SATA drive breaks the 100 MBps barrier! And yes I know the difference between maxium transfer rates and average rates. There is currently a 15000 RPM SCSI drive that peaks slightly above the 100 MBps barrier. Have you ever looked at any hard drive benchmarks?" Haven't disagreed with this, why you use RAID! My god man, how many time has this got to be discussed?

"And before someone else misreads a benchmark and brings this up. "
Seems you keep your eyes closed during test readings. When you need throughput, you use RAID. Why you refuse to see the advantages of RAID is beyond me.
I'm now putting you in the category of the ignornat that spread ignorance.
RAID isn't a magical HD steroid but when used for what they were intended for, they help tremendously. You stick with 1 HD and suffer along, I'll keep my RAID and save time and money. Once again, if RAID sucked monkey nuts like you seem to want to say, why is it used in corporate setting every day, 24/7? Just for data security? Why is it I (and thousand others) see noticable, measurable real world gains by using RAID? If it was a wicked trick and slight of hand, I'd say so, but it isn't. Why I have dome many tests on my own. I have a few 40 GB drives laying around, maybe I should send them to you so you can test for yourself. Even just 2 drives in RAID0 will break your magical 100 MB barrier. Open your eyes. As far as your misinformation posts on RAID, I will no longer reply as you are being silly. If you spread misinformation long enough, will it become fact? You seem to think so. Good luck with your crusade to bash RAID, cheers!
May 1, 2006 1:47:02 AM

You are very confused about something. I don't know what. Maybe its the english language, maybe its about how forums work.

You see someone create a topic then people repsond to that topic and respond to the responses. If someone want to sudeenly talk about something completely different they start their own topic.

This topic of this thread is the myth SATA II devices.
SATA II is not RAID
SATA II and RAID have almost nothing to do with each other.
RAID has no relevance to our conversation.
No one here is discussing RAID.
No one here is attacking RAID.
No one here is discussing ways to improve performance.

At first I though you were trying to make some point about RAID techniques having relevance to our conversation.

I then mistakenly concluded that you were confused about RAID. I realize know that you are not confused about RAID. I appologize.

I now realize that you were just interjecting comments about how RAID can be usefull in increating performance.

Had I know this I would have simply agreed that RAID can sometime work wonders.

Instead I replied by explaining by pointing out that RAID techniques do not have an impact on my argument "STAT II does not exist".

The fact that RAID has nothing to do with out conversation here doesn't me that I don't like it or that it is bad. I like many things that have nothing to do with this converstation.

Please do me a favor and quit posting defending RAID".

If you want to continue talking about RAID please start your own topic.

Maybe you can call it "RAID does not suck Monkey Nuts", and you can explain to everyone about how great RAID is and how it does not suck the nuts of monkeys. I will then gladly agree with you that RAID 1 is wonderful and even RAID 0 is best choice for some situations. I may however argue that RAID 0 without a sound backup strategy could result in some form of testicular suction invovling a donkey.
a b G Storage
May 1, 2006 2:30:56 AM

I do agree with the subject line but am gonna post anyway...but not about SATA hard drives...

I love the guy sat work who are now braggin about there computers with SATA optical drives...hahahahahaha!

I usually just nod and smile and say how cool that is and then zone out when they go on and on about how much faster it reads cd's, dvd's, and blah, blah, blah...
May 1, 2006 3:11:34 AM

Well then as punishment for going off topic I will go off topic even further.
---

Wow at SATA 150 MBps speeds, with 1x DVD = 1350 KBps they must be getting between 75x - 113x!

My 16x ATA burner is a peice of crap.

--

You know I know this expert on home theater equpment who keeps explaing to me at length how music sounds better over a digital connection uses an coax cable instead of the optical.

Its because the different frequencies of light travelling along the fiber optic cable end up travelling at different rates.

Countered by explaining how my digital photo's always look better when I email them over analog connection instead of DSL.

In all seriousness he suggested I switch to a DSL provider with dialup access.

Another time he lamented the fact that his TV didn't support HDMI because he wanted to connect his PC to his HDTV and take high definition screen shots of movies.

The scary part is this guy knowledge of Home Theater equipment is encyclopedic. He can tell me what DLP TVs support HDCP, which support HDMI, compare and receivers in the $1500-$3000 range you name it.

But one level his is very compenent and an expert yet some deep level he really doesn't "get-it". He definately knows the facts, he definately knows how to get the job done when it come to anything home theater related, but on the other hand he doesn't get the basic concept of digital vs analog transmission.
May 1, 2006 6:42:34 PM

Quote:

2) The transfer rate is for the link between each individual drive and the controller. For PATA that link may be shared by a master and a slave. For SATA one drive gets the whole 150Mbps-300Mbps.


I have to disagree there, because some motherboards use a port multiplier to get the 4 SATA ports. In that case, raising the bandwidth rate does matter. For example, 4 drives that are RAID 0'd can take benefit from the 300MB/s bandwidth rate, but at 150MB/s they would end up maxing out.

For ports that are allocated single SATA channels for each one, then you are correct, absolutely no benefit. If you made that distinction, sorry I missed it.
May 1, 2006 7:14:58 PM

There IS a benefit to a 300 Mbps SATA connection over a 150 Mbps SATA connection... it's just a VERY tiny one... a hard drive's cache would fully utilize the additional bandwidth of a 300 Mbps interface. Most desktop drives have 8 MB (or less) of cache so this isn't a huge benefit... but it IS a benefit.
May 1, 2006 7:31:31 PM

I have an Hitachi 250 Gb with 8 Mb. It was factory set on SATA-I. I plugged it on a Nforce430 controller, and tried setting up NCQ. Until I switched both pieces of hardware to SATA-II, I couldn't get it to use NCQ.

Once there, there was a VERY slight increase in transfert performance - not much, but probably faster access to the 8 Mb of cache helped. However NCQ has a damn effect on reading small files - indexing a drive takes much less time now.

SATA-I defined a serial medium for ATA protocol. SATA-II defines an increased frequency AND adds a feature. True, a drive isn't SATA-II if it doesn't support NCQ. If it support both 3GHz and NCQ, then it is SATA-II.

What was the difference between USB 1.1 and 2.0? They are 100% compatible, only USB 2 uses 480 Mbps while USB 1.1 is limited to 12 Mbps.

What was the difference between AGP and AGP 2X? Dual signal yes, but also sideband addressing; AGP 4X? Fast Writes. AGP 8X? None.

So SATA-II is double the transfert rate (not much use with existing drives yet, but in the coming 1-2 years its bandwidth may not be enough anymore), with an added nicety: NCQ.

Sorry if I missed some things, I read the thread very fast and may have missed someone saying the same thing.
May 2, 2006 4:59:17 AM

I know what you mean, I like them just for the small cables and that is about all I can see in using them for. Ohhhh, not faggots but SATA HD's, Oh yea, this is no place to bash people Let It Go! :wink:
May 3, 2006 1:54:08 AM

I said having a 3.0 Gbps controller is best.

Its the drives where the two transfer rates make no practical difference.

And making faster access to the cache doesn't make any practical difference.

Anything on the hard drives cache will be in the OSes cache for that drive using memory thats over 80x faster.

That cache if for the drives internal use. Mostly so it has a place to put stuff so it doesn't have to operate strickly FIFO.

Bottom line SATA-IO says never use the labels SATA I and SATA II to describe products and when it comes to SATA their word is final. SATA 150 and SATA 3.0 Gbps are NOT like USB 1.1 and USB2, and finally never assume a drive with a SATA 3.0 Gbps transfer rate is automatically superior to a SATA 150.
May 3, 2006 3:00:27 AM

Santa Claus is fake.

"SATA-II" Adds greater functionality over 1st generation.

Better hot-swapping performance, by far.
Port Replication.
Longer cabling.
Better error correction. (CRC)
More headroom for burst speeds.

As for people crying about 300MBp/s theoretical limit. I guarantee you I can saturate that easily.
Silicon Image. Sil3132 + Sil3726 in a RAID-5 or nested 0+1. Can you say 5 HDD's at 60MB/s thoroughput? 5 x 60 = ?? OMGWTFBBQ 300!!

NCQ mainly helps in databasing of large files, dampening the effects of fragmentation, and accessing a large paging file.

For your standard desktop user and web surfer, SATA Gen 2 will not make a big difference in the slightest. For those of use that actually use computers for something besides wang extentions, it can potentially help alot.

Perpenicular HDD's will help increase data density and size, as well as help with thoroughout. Eventually. It's still first generation, barely at that.

P.S. USB 1.0 Low Speed is 1.5Mbp/s, USB 1.1/2.0 Full Speed is 12Mbp/s, USB 2.0 HiSpeed is 480 Mbp/s.

Flame accordingly if you object with common sense and facts, I don't care.
May 3, 2006 1:25:24 PM

Am I not correct in thinking that SATA theoretically allows for 'hub' type devices, just as SAS does?

If so, that would be a theoretical advantage, assuming you can find a 'hub' then it becomes alot easier to saturate bandwidth

EDIT:realised thats what the aforementioned Sil3726 is. They are a bad idea imho and I'd be willing to bet alot of controllers wouldnt even support them, but its an example of a use for SATAII
a b G Storage
May 3, 2006 1:41:46 PM

KWH, quiet please. I am not accusing you of ignorance, but perhaps of misunderstanding. Codesmith is discussing individual drive performance and features as they relate to the SATA specs. The performance of a RAID is not relevant to the discussion.
May 4, 2006 4:15:28 AM

I can handle the blunt truths as they relate to computers, but to so coldly come right out and say Santa Claus is fake right out of the blue, well it's just not right and now I'll likely need expensive therapy to get over this. Is it really true? Is Santa really fake? Perhaps he's not, he's just Rambus.
May 5, 2006 12:27:49 PM

Hmmmmmm I sure feel stupid after my new post about Sata II.

;-)

Oh well, I honestly didn't know and I'm a noob.
!