SATA 1.5 Gb/s vs. 3.0 Gb/s

Ok.

I've been researching this for days and I'm finally giving up and I need to find out once and for all:

Is there ANY performance increase by going from a SATA-150 to a SATA-300 HDD? At all?

Here's the thing... I just purchased an Inspiron 1720 notebook. I want to replace this factory mainstream SATA HDD with 2 fast SATA HDDs.

And just so it's out there... The new Inspirons DO have SATA-300 capable controllers. My question is more regarding if doubling the bandwidth from 1.5 Gb/s to 3.0 Gb/s is beneficial at all.

I've heard that the extra bandwidth doesn't matter, as hard drives - especially notebook drives - don't even fill up the 1.5 Gb/s that SATA-150 is rated to provide.

Alternatively, I've read that if you have a SATA-300 capable HDD, you can still read and write from the HDD's cache at the full 3.0 Gb/s... improving burst rates and any long process which fills the cache with small enough chunks as it's writing or whatever...

Let's take Hitachi's Travelstar 7K200 drives. Some of them are SATA-150, some are SATA-300. That's really the only difference between the 2 sub-series; everything else is the same - including the 16 MB cache.

So, even though the throughput of these SATA-300 drives will never actually reach 3.0 Gb/s... (or even 1.5 Gb/s as I understand it... even in RAID) ...would there not still be a slight performance increase by doubling the bandwidth for at least the 16 MB of cache? Also... is there anything else regarding performance that I'm overlooking?

Don't tell me that SATA-300 sucks more power. I know that.



Thanks!

 

http://en.wikipedia.org/wiki/Serial_ATA

wiki is your friend

 

You answered your own question. Just look at reviews and they will tell the story. Infact, SATA-150 - SATA-300 - and even PATA doesn't really matter, interface is really not the botteneck, it's actually the drive technology that matters.

 

The only thing SATA-300 will increase is your power consumption. But you knew that. So yeah, no improvements.

 

Well the thing is it makes no difference for YOUR situation, however it does matter for RAID arrays and RAM drives, so there is still a utility for it, just not so much for laptops AT THIS TIME.

Dont' forget CF's next revision is for around 300MB/s troughput which gets you past the 1.5Gb/s barrier pretty easily. So if they move that new standard to SDDs you may see utility for in in the future. Imagine storage memory faster than the system memory of just a few years ago.

 

Actually, notebook RAID with two drives won't be affected much either, if that.

 

I wasn't talking about notebook raid, nor with just 2 drives.

 

Touché. ninjatextisback

 

Word. Thanks dudes.

 

BTW, look up the specs on CFast and then think about that technology in an SSD (likely to be adopted shortly thereafter) 300-400MB/s from a single CF unit.

There's you reason for SATA 3G in a laptop in just over a year's time.

For now like D3X mentions the sticker on those same old magnetic drives doesn't matter much because individually they can't do it on their own.

 

Right, well future SSDs will undoubtedly be very interesting. On paper and theoretically it should be way faster, but it all comes down to "how much it's gonna cost for the amount of storage" and how that translates for the average consumer.

My graph shows what's out there presently and not talking about future technology, which often is really just a cloud of a vision...

 

Yes, but that's always been the case with storage with everyone thinking there's magical barriers and it will be too expensive to do X, until that next barrier is easily surpassed and we have 16GB CF/SD cards for about the price I paid for highspeed 256MB just a couple of years ago. Same with magnetic drives.

If this technology is aimied at Cameras and ever large sizes for solid state HD camcorders as well, it's not going to be expensive for long and even with a 16GB CFast drive I'd make my it primary drive this instant and move my current drive into the next bay for a dump drive.

And no your graph doesn't show all of what's out there presently, RAM drives are much faster, even the fairly cheap but small iRAM drive already was at the limitations of the 150 interface and beyond the PATA limits.

So I don't see how this is so much a cloudy vision, versus if I were talking about next gen holographic storage. It's just over a year away from mass production so it's not some pipe dream.

 

Right, but CF drives at the moment only top off at 300X which is approximately 40MB/sec so I don't see how these would be on average faster than the traditional magnetic drives. I have yet seen anything larger and faster for a 16GB or 32GB that would be beyond the 300x mark on sale yet The fastest I've seen is at 166X which comes at a very high price point.

I don't see how iRAM would be even considered as "Fairly cheap", The Gigabyte device utilizes desktop ram and creates a storage device. It's definitely innovative but this is beyond being cheap even at today's ram prices, you would be spending nearly $400($160 for the iRAM card alone, 220 for 4x2GB) for a mere 8GB total, that's a nearly $50/per GB. Compared to WD Scorpio 320GB at $180 which works out to be $0.56/per GB. What also doesn't make sense from your statement is that it would be faster than SATA-150, well the iRAM doesn't even use SATA, infact a desktop PCi-e slot which much higher bandwidth from both the memory modules as well as the buspeeds. We're talking about notebooks here(no duh, it's NBR), so how iRAM is related to this is beyond my comprehesion. You might as well start mentioning enterprise level Fibre channel/Infiniband devices...

I'm not saying that SSDs won't be the future, it definately is, but presently even the fastest SSD made by MTRON 32GB (reaches only the 120MBps read/90MBps write), and that achievable even on all interfaces including PATA/EIDE/UDMA-6.

I believe my graph shows what's out there "within reasonable monetary value" as a viable solution for the average consumer. I wouldn't be suggesting MTRON SSDs 128GB for a measely $1600 for everyone, now would i? If you can afford it, get it, but even still I won't think it's worth every penny.

 

It not on sale right now, but the CFast cards are demoing right now. It's like when you see the pre-release AMD, intel, nV parts at CES, CeBit, etc. They are currently just demo products, but it is what will be production in about 12-18 mths. So it's not crystal ball gazzing, it's actual physical items, just currently in standards phase.

It's considered 'fairly cheap' compared to the other commerical RAM drive solutions out there, but it was a statement about the interface not the perceived 'value' to someone else.

And how much are far slower SSDs right now?
16GBs can start at just over $400 too. The iRAM drive is significantly faster in all areas, but like I said limited in size. However, make the 16GB SSDs out there perform like the iRAM and I'd happily pay twice the price @ that $50/GB no problem, as would many other mobile photo/video editors, because finally they'd be worth for the speed alone without the drawbacks of RAID in laptops.

Wow you didn't even bother to read the article you linked to did you?
I know the pictures are deceiving, but read the article it's SATA, the PCI slot (not PCI-e) just supplies power to the DIMMs and rechargeable battery so that it keeps the info after power down.

"Then Gigabyte unveiled the i-RAM, a $150 solid state-storage device that plugs directly into a motherboard's Serial ATA port"

"And then there's the Serial ATA controller's lack of support for 300MB/s transfer rates, which will probably be the card's most serious performance impediment."

Well if you don't read what people write, I understand completely why you have a comprehension issue, of how CFast style SSDs would benifit notebooks. (edit: It's NBR man, why would you want lower power consumption and super fast, Duh!)

Seriously, I said to the OP that he currently has no major use for SATA-300 right now, however he will shortly, and that the myth 'that even SATA-150 isn't saturated yet' is perpetuated by people who don't know WTF they talking about, probably posting graphs to back up their statements.

I believe your graph is irrelevant to what I was saying, which didn't contradict what you said initially, in fact I supported your original statement. You just can't seem to figure that out or understand why I mentioned faster drives in the first place.

I've said it basically 3 times sofar, it's not about needing it right now, but very soon it will come in handy having that faster controller, perhaps even in the lifetime of him owning this laptop he's talking about.

And really no one cares if you think they're worth every penny, no ones looking for your approval, these things aren't for you obviously. Hope none of the SLi owners have to listen to you go on about being 'worth every penny'.

Seriously, you'll get much further in life if you read, instead of just looking at the pictures and graphs.

 

Ugh this is getting old real quick, and I wasn't trying to pick a fight here. Comprehension is based on the information relative to the original topic?

Right, I must have missed the part with the SATA controller on the iRAM but seriously as the concept of iRAM is far and few between to be applied to notebooks and obviously not a viable option(that thing needs a battery as well and once it's out your data goes "poof").

You just don't seem to get what I'm aiming after, granted that SLi and AW users are paying top dollar for the products, but how much of percentage of users would that relate to compared with the rest of the market?

Enough about this future technology, as this is pointless. Besides, you can't believe everything you read either.

Anyways getting back to the point of this thread was for the OP purchasing decision for his m1720 (just incase you didn't read it), and if you really analyzed what I was trying to imply was to either go with Hitachi 7K200 200GB or the WD Scorpio 320GB which is at the same price point that displays both SATA-150 and SATA 300 ironically but performance wise are pretty close and are pretty much at the top of the list excluding SSDs.

 

So here's the thing, do like I suggested to the OP and type CFast into your google/yahoo/whatever search bar and look at the specs and some of the recent product demos from CeBit. Then think about the following: FASTER than that iRAM for transfers (almost definitely not for random access), plus, small low power and non-volatile. Perfect for Laptops and voila, reason to be glad his laptop has support for the higher transfer rate. That was my point, not that there's a need now, but soon very nice benefit.

What an interesting thing to say on a text forum.
The thing is they are already demoing it. I trust that Montevina is coming, the RV770 is coming, the G200/T100, and a bunch of other items are coming because they aren't outlandish. What I never bought was people saying OLED laptops before the middle of the decade, but by the end of the decade heading into the next one, sure.

See and that's the thing, I wasn't analyzing what you said, cause I didn't disagree with it, I was replying to both fabarati and therealmatt's statements. Specifically the OP's statement "The new Inspirons DO have SATA-300 capable controllers. My question is more regarding if doubling the bandwidth from 1.5 Gb/s to 3.0 Gb/s is beneficial at all."

I think if you look hard at what I wrote you'll see what I said is "right now it's of little benefit (because as you mention they're currently limited by the properties of the magnetic drives), however soon they will show their beneficial side with newer drives. Desktops can use the extra throughput under certain situations which I illustrated, just not in laptops at this time"

I thought I was pretty adamant about that twice when I mentioned it to be sure not to step over what else was said in the thread, obviously that went un-noticed.

So regardless of nay-sayers; I still say ever increasingly fast SSDs are the main reason to be happy if you have a faster SATA controller, because perhaps before you upgrade to your next laptop you'll want to add one of those faster SSDs even before they reach CFast speeds.

 

Jeez, cool down. No, CFast is not going to magically revolutionize laptops a year from now. First, the most recent source I could find had manufacturers saying that it might start showing up 18-24 months from now. Then add another 6 months (at least) before it gets adopted into laptops, and we're talking *at least* two years. And that's assuming it'll be worth using the very first models (which probably won't be the case, because of price and capacity)
Then there's price. Flash ram is expensive. Flash RAM isn't magically going to be vastly cheaper next year.
And capacity. Right now, we can get what, 16GB SSD drives... Or 500GB harddrives.

And finally, most obvious: Until the product is on the market, nothing is certain. it might get cancelled tomorrow, it might disappoint performance-wise, it might not be offered as a standalone harddrive replacement.

Er, no. It's not saturated *yet* because your hypothetical CFast is not out *yet*.
As much as this may surprise you, a bus can not be saturated *today* by a product that's out *tomorrow*.

No. Laptops need two more factors.
Price and capacity. A standard that's developed for cameras *may* not provide sufficiently high capacity to be worthwhile as a harddrive replacement, *and* it may not reach anywhere near an acceptable price per gigabyte.

And finally, this has nothing to do with what the OP asked. He asked if there was an advantage to buying a SATA 3.0GB/s harddrive instead of a SATA 31.5GB/s.
Nothing about the interface. Of course his motherboard supports 3.0Gb/s already, otherwise he wouldn't ask. The question is whether there will be a difference in the performance of these two *harddrives*.

Ranting on about how he may want to buy a computer with a chipset that supports a technology that may be available 2 years from now, and may be worth using is completely missing the point. He's got the computer, and he's got the chipset, and it supports SATA 300.

 

w00t, Jalf to the rescue! Tell it like it is. Other clichés adapted to this particular happenstance!

After G****Ape started ranting, I didn't wanna touch this thread with a ten foot pole, even to reply to his statements about me (and the OP).

 

Wow Jalf, I'm surprised you read all of that! Thanks for summing up what I've been trying to say... +REP!

 

Tell me about it, I fell in and couldn't get back up

 

yeah, sorry about that. I went with a basic pronciple of the internet: don't feed the troll.

 

Wow you guys really are myopic and reactionary.
I'll grant I'm being defensive, but in light of your arrogance compounding your ignorance, I think I'm pretty justified.

It's not magic, FLASH ram is historically cheaper every refresh, just like most computer products. It's a question of economics and principles like Moore's Law.

The reason I use CFast as the example is because that's the point at which the underlying NAND technology makes sense as a consumer solution and not one limited to servers and workstations.

As for the sizes, two things: first even if they shipped today in size CF currently sells at they would be sufficient for most people as a primary drive with their slow drive in bay 2; second SSDs aren't limited to 16GB, they already sell 128GB in the consumer segment and the are just starting to sell 512GB in the enterprise market. SSDs have increased their capacity far quicker than magnetic drives.

OK, so let me just dispell that 'most' obvious limitation you confuse. They already make and sell 200MB/s 3.5 inch drives which already saturate the bandwiddth of SATA-150, the only question is how long to move it into the notebook level from enterprise desktop.

Yes, it's already saturated by current SSD products and was saturated years ago by consumer products like the iRAM listed above. It's only a question of time before it's saturated on notebooks, and I'm not talking about products based on the NAND CFast is made from I'm talking about the NAND modules currently ramping up production for product for Xmas. Even if delayed that's saturating the bandwidth in the lifetime of his laptop.

Laptops don't need price, that a function of the buyer/seller, not the laptop. Capacity is a relative thing, like I said, some people would prefer an ultra fast 16-32GB primary drive to speed up apps and load times and have a large dump drive for the stuff that's not that dependant on bandwidth.

The underlying NAND standard is pretty applicable to all products, it's not about the form factor, which seems to have stuck you all.

I wasn't ranting, but you guys are attacking, and you do totally miss the point. Obviously I got the point when even in my initial posts I supported D3X's statements about the interface for those drives, but I simply went on to add though that the OP's other questions have a different answer you guys either ignored or just didn't understand.

And while it currently has little benefit, he'll get a nice free boost when SSD performance improves so be happy and thankful he's got it. That should be simple enough to understand I would think.

 

Actually you jumped in and realized you we're in over your head.

Like I said, read my second post it supported what you said while adding that there is future benifit.

Seriously after you guys are done patting yourselves on the back, open your eyes and stop being so ignorant.

 

We're not being arrogant about our ignorance. We said that HDD's and notebook raid don't saturate the Sata-150 bandwith. We never said it won't do that in the future, just that it doesn't. We meant now. Notice the present tense. We never talked about SSD's either. We just answered the OP's question. Which was: Is there a benifit to getting a sata-300 HDD over a sata-150. And the answer (right now) is No, there isn't.

 

Correction, I was not over my head. The only thing that's over, is this topic and this thread, as it has gone way over what was supposedly discussed as OP's concern. Notice he has not responded since? And I can easily say that the Ape rant monster has scared him away...

I understand your optimism with the future technology and that too gets me excited as well. However with my experience and the knowledge that I have and how many years I've been following the industry, your optimism seems a little too optimistic. I can easily say that majority of those products will never reach the consumer level (perhaps maybe CFast) within the next 2 years it will likely be 3 to 4 years from now before it even hits mainstream. Mainstream meaning that it will appear as a standard product avaialble at your local Best Buy.

 

Best Buy doesn't currently sell SSDs (except in their Business/Enterprise section) so that's a tougher standard than most people here would use.

As for being optimistic, nah, I've seen things that never made it to the public thanks to where I work, however when a technology is simple as NAND it's not like there's some giant leap of faith to think that if it's in a CF card for cameras and digicams it'll be incorporated into the consumer SSDs by that time.

Seriously considering the full circle coming back to what I originally said I don't even understand why you bothered challenging it. I didn't say anything revoutionary and even tempered it with "you may see utility for in in the future", that's not telling him to wait or not to buy or to buy something specific either, just hey, but here's a nice free bonus MAYBE.