AMD, Seagate demo SATA 6Gbps

[LiamC]

..."Despite its fatter pipe, SATA 6Gbps will use the same connectors as the existing Serial ATA standard. The spec includes new goodness beyond a higher throughput peak, too, but AMD and Seagate are only talking data rates for now."...

http://www.techreport.com/discussions.x/16544

 

[P5-133XL]

Using the same connectors is all nice and fine, but more importantly, it needs to use the same cables. Does it do that?

 

[udaman]

Using the same connectors is all nice and fine, but more importantly, it needs to use the same cables. Does it do that?

Yes.

ESATAp is more interesting to me, along with the dual USB2.0/ESATA port. Think I read somewhere, Intel's current chipsets don't support the standard yet.

http://www.pcworld.com/article/156535/faster_serial_ata_coming_this_year.html

A big growth area for E-SATA: 85 percent of new laptop PCs gained E-SATA in the second half of 2008, thanks to the use of a clever two-in-one USB 2.0/E-SATA port. This port didn't originate from the SATA-IO standards group, but has clearly gained significant traction.


The third -generation spec for SATA doubles the speed from 3Gbps to 6Gbps, says Mawell. Products and the full spec are coming in the second quarter of 2009.


The intention is to double the E-SATA spec, too, to 6Gbps.
"This will help consumers transmit data faster," says Mawell. "It's backward compatible with SATA-1.5Gbps and SATA-3Gbps same connector and cable as before, so it will automatically adjust the speed based on the device support."


SATA 6Gbps drives will be shooting for performance that exceeds that of USB 3.0 and of Solid State Drives.


Also coming with the new spec: Power over ESATA (ESATAp), which allows 5Volts or 12Volts of power to be delivered via the ESATA port. Maxwell expects we'll see SATA 6Gbps appear in enterprise and SSD storage first, then see it trickle down to performance PCs and notebooks.

 

[udaman]

And in the USB3 corner:

http://www.pcworld.com/article/156494/superspeed_usb.html?

Compatability

Backward compatibility shows up in the new standard. USB 3.0 maintains the extensive device class driver infrastructure of USB 2.0, and a USB 2.0 devices will work via a 3.0 connector.
Performance

Performance improvements are notable, too. "We have research that shows that after 1 minute, 1.5 minutes waiting for a transaction, users get impatient.," says Ravencraft. "The transfer times have to get much faster." Hence the evolution of USB SuperSpeed's Sync & Go concept.
SuperSpeed USB 3.0's 5Gbps data rate (compared with Hi-Speed USB 2.0's 480Mbps) should help--and with very clear real-world advantages. For example: A 25GB HD movie will take 13.9 min to transfer with USB 2.0, and 70 seconds with USB 3.0, says Ravencraft.


Practically speaking, the implications are tremendous. Imagine not having to wait hours on end for your full-drive data backup to complete, or not having a lengthy delay when off-loading 32GB flash memory cards from your digital camera.
Power

As for power consumption, "power today is king for portable devices. It is the pinnacle, the focus for the PC, the notebook, and all devices. All of these new specs had to be optimized for power efficiency," says Ravencraft.


For example, there's no longer any device polling, so connected USB devices can enter a virtual sleep mode; more power will be able to go to the device (which will hopefully eliminate some of the power issues we see today with portable hard drives that require extra power from a second USB port); and USB 3.0 will no longer broadcast information to all connected devices, thereby saving power, too. Plus, when a device's battery is drained, it will now still be recognized by a laptop, for example, so you can charge it (this doesn't work with USB 2.0).


The intention is for SuperSpeed USB 3.0 to provide headroom for the next five years.


"We've designed the protocol so it should operate at up to 25 gigabit per second data rates if the pipe supplies that type of data rate," says Ravencraft. "I don't think we'll get to that over a copper wire, but we might look at an optical interface next. And we're ready if it does."


Ravencraft expects host controllers and device controllers by mid-2009, and consumer products by early 2010. His estimates are more conservative than those offered by Symwave, which is demonstrating USB 3.0 at CES 2009 in conjunction with Seagate.

Hmm, 70sec for 25GB, means it will still take 1hr and 21 min. to backup up a full 1.5TB drive. A faster I/O right *now* wouldn't hurt, 25Gbps would be nice...could clone a 1.5TB drive in 17min...now yer talking!

 

[Mercutio]

Using the same connectors is all nice and fine, but more importantly, it needs to use the same cables. Does it do that?

It's a serial interface. One wire to send, one wire to receive. I am thinking it will.

 

[P5-133XL]

It's a serial interface. One wire to send, one wire to receive. I am thinking it will.

In theory, as the frequency goes up all sorts of bad things can start happening, even on a serial cable: Impediance changes causing connector reflections, crosstalk can start happening. Really the only thing that doesn't happen on a serial interface as opposed to a parallel is a race condition.

In actually, I've now seen several other articles that actually specificy that the cable is the same too. So everyting seems to be good in that it will be backwards compatible to the 1.5 Gbps and 3 Gbps standards.

What I'm surprised about is that there wasn't a bigger jump considering the fact that SSD's should shortly be limited by even this new standard.

 

[ddrueding]

Some of the beast SSDs are already passing 768MB/s and are using PCI-E directly instead. Might we see a PCI-E cable connection? No doubt lengths would be extremely short.

 

[P5-133XL]

I also wonder if this is the start of the seperation of SAS from SATA. Before this, SATA was a subset of SAS so that you could plug a SATA drive into a SAS controller. However, I haven't seen any reference to SAS running at 6Gbps ...

 

[ddrueding]

I wonder if this is the end of SAS? It was supposed to be the high-end, but if SATA beats it on speed, and backplanes and RAID controllers with 24 ports negate the multiple-drive-on-a-cable thing, what else does it have going for it?

 

[LunarMist]

I thought 6Gbps SAS was announced around 6 months ago. Why abandon it?

 

[ddrueding]

Hadn't heard of that...I'll eat my words.

 

[LunarMist]

How long will it be before there are SATA 3 RAID controllers?

 

[LiamC]

Sorry in advance. Geek humour:

Posted by Travelmug over on TR

..."So the guys who can't make a working HDD team up with the guys who can't make a working SATA controller for their chipset in order to demonstrate the latest SATA spec. This should end well..."...

 

[LiamC]

Sorry in advance. Geek humour:

Posted by Travelmug over on TR

..."So the guys who can't make a working HDD team up with the guys who can't make a working SATA controller for their chipset in order to demonstrate the latest SATA spec. This should end well..."...

 

[ddrueding]

That is funny, Liam.

 

[P5-133XL]

A MB with a 6Gbps SAS2 controler onboard.

 

[sechs]

I thought 6Gbps SAS was announced around 6 months ago. Why abandon it?

http://www.scsita.org/aboutscsi/sas/SAS_roadmap2004.html

The 6GB/s plugfest was in November.

 

[LunarMist]

They were announced in September, so that is close enough for me.

 

[udaman]

SATA 3.0 specs finalized

When will the SATA 6Gb/s SSD's be shipping?

http://www.sata-io.org/technology/6Gbdetails.asp

The SATA Revision 3.0 Specification enhancements include:

A new Native Command Queuing (NCQ) streaming command to enable isochronous data transfers for bandwidth-hungry audio and video applications

An NCQ Management feature that helps optimize performance by enabling host processing and management of outstanding NCQ commands

Improved power management capabilities

A small Low Insertion Force (LIF) connector for more compact 1.8-inch storage devices

A connector designed to accommodate 7mm optical disk drives for thinner and lighter notebooks

Alignment with the INCITS ATA8-ACS standard

...and many more!

 

[LiamC]

The important part is:

..."the technology itself is to be referred to as "SATA 6 Gbit/s". A product using this standard should be called the "SATA 6 Gbit/s [product name]".

The terms "SATA III" or "SATA 3.0", which are considered to cause confusion among consumers, should not be used."...

emphasis mine.

 

[Pradeep]

http://www.seagate.com/ww/v/index.j...toid=8dda75f6f8472210VgnVCM1000001a48090aRCRD

Seagate 15k.7 (600GB 15K) releasing in 6Gb/s SAS and 4Gb/s FC

 

[Mercutio]

For reference, attaining 750MB/sec data rates would require 24 total PCIe lanes. PCI-X tops out at IIRC 4Gb/sec. I've never even seen anything with an Infiniband interface. I know they exist, but that's the only thing I know that could keep up with a 6Gb/sec connection.

 

[sdbardwick]

Each PCIe lane is 250Mbyte/Sec (1.1) or 500MByte/sec (2.0). PCIe FAQ link.

 

[udaman]

For reference, attaining 750MB/sec data rates would require 24 total PCIe lanes. PCI-X tops out at IIRC 4Gb/sec. I've never even seen anything with an Infiniband interface. I know they exist, but that's the only thing I know that could keep up with a 6Gb/sec connection.

And what do you think Joon is using to edit 4k RedOne captured pr0n with Merc?

HP has an infiniband Raid solution, but there are more with mini-SAS interface...see below. According the BF, newegg had the 4322 on sale @1/2 price in March for $339, it's expensive. But you could still get a 2314 and 4 SSD's and do it for less.

RocketRAID 4322

http://www.barefeats.com/hard109.html

uh, hp announced a 1.2GB/s Raid controller in '07 Merc, old news...bet you could max that out with Vertex Turbo series:

http://www.tweaktown.com/news/8295/highpoint_announce_1_2gb_sec_sata_raid/index.html


[FONT=Verdana, Arial, Helvetica, sans-serif]HighPoint eSATA for Mac[/FONT]
http://www.barefeats.com/hard119.html

INSIGHTS
We found the HighPoint RocketRAID "eSATA for Mac" to be a truly versatile SATA host adapter. It works with Port Multiplication (PM) and Direct Access (DA) enclosures. It works with "RAID-in-the-box" and non-RAID enclosures. It works with single bay and multi-bay enclosures. If your external storage device doesn't have built-in RAID functions, you can use HighPoint's brower-based RAID manager to define RAID 0, 5 or JBOD. Finally, you can boot OS X from storage devices attached to the card.​

The maximum throughput per port is 300MB/s. Certain storage solutions are faster than others by nature. The fastest setup was when we striped (RAID 0) two OCZ 120GB SATA 2 Vertex SSDs, each with its own eSATA port, attaining 400+ MB/s. The maximum theoretical throughput of the "eSATA for Mac" is 800MB/s. We have actually achieved that speed using four SSDs on the RocketRAID 2314, which uses the same chipset as the "eSATA for Mac" but has 4 ports.​

Port Multiplier (PM) devices are limited to around 230MB/s no matter what drives you use or how many of them you have. The FirmTek SeriTek/5PM is a good example of a PM enclosure. When we installed two fast SATA drives in a RAID 0 set, it reached the 230MB/s limit. When we bumped it up to three drives, the transfer speed did not increase, but neither did it drop when the RAID 0 volume filled up. The other PM enclosure we tested with the new Oxford 936QSE or 936DS chipset (Wiebetech RTX400-QR) ran slower than the SeriTek for some reason. We're still researching why that's the case.​

If you want more speed than what we achieved with the dual bay Direct Attached or PM enclosures, you'll need a card with either four eSATA ports or a mini-SAS port and possibly a stronger I/O processor -- like the RocketRAID 2314, 3522 or 4322. You'll also want a Direct Attached (instead of PM) enclosure with more than two bays -- each drive having its own dedicated port (aka 300MB/s data channel). But if you just want an easy-to-use, affordable way to connect external SATA enclosures to your Mac Pro, the RocketRAID "eSATA for Mac" certainly gets the job done.​

 

[Mercutio]

I wasn't aware that PCIe had been revised.

 

[sechs]

Maybe that's because they didn't do anything goofy like change the physical interface or add any badly needed features.