Slashdot Mirror
USB 3.0 the Real Deal, SATA 6GB Not Yet
MojoKid writes "HotHardware has posted a sneak peek at a new motherboard Asus has coming down the pipe with USB 3.0 and SATA 6G support. The Asus P7P55D-E Premium has a PLX PCI Express Gen 2 switch implementation that connects to NEC USB 3.0 and Marvell SATA 6G controller chips. With a USB 3.0 enabled external hard drive connected to a USB 2.0 port and then to the board's USB 3.0 port, there were some rather impressive gains to observe. When connected to a USB 3.0 port, the external hard drive was about 5 — 6x faster versus connecting over USB 2.0, with total throughput in excess of 130MB/sec. On the other hand, benchmarks with Seagate's new Barracuda XT SATA 6G drive show little performance difference but a burst rate that is off the charts. According to ATTO, there are slight overall performance benefits to be had connecting the drive to the SATA 6G controller, but the deltas were quite small; somewhere in the neighborhood of 5MB/s or so."
oh this sounds too good to be true
So what is there in usb 3.0 that we did not already have with firewire?
Speed, price and ubiquity. HTH. HAND.
The Barracuda XT is a spinning platter HDD and so should not be expected to benefit significantly from the new SATA revision. SSDs on the other hand have already maxed out the transfer rate SATA 3Gbps. I suspect they would have seen the difference if they used a top of the line SSD.
This is good news all around, it's great to see things getting faster.
Make the chart bigger!
This all sounds like exactly what you'd expect.
The old SATA standard was more than sufficient for the hard disk's max sustained transfer rate, so only burst performance (when everything is presumably coming from the disk's RAM cache) changed with the new SATA. So "SATA 6GB" is working fine, but this disk is just too slow to take advantage of its speed increase.
With USB on the other hand, USB 2 is simply far too slow to handle even the drive's sustained transfer rate, whereas USB 3 is fast enough to handle it.
So the moral seems to be: USB 2 sucks for disks, USB 3 is better and probably sufficient for a typical hard drive, and SATA's still probably better than either (it's not really possible to tell from this article, since the sustained transfer rates are limited by the drive, and they curiously omitted the burst rates for USB).
We live, as we dream -- alone....
SATA 2 is already a bottleneck for many SSDs as this chart shows them hitting a wall at approximately 260MB/s. SATA 3 should release the proverbial floodgates for sequential reads.
On a tangent, Samsung just started mass production of a 64MB, 60nm phase-change RAM in September. Initially they are going to use them in mobile phones. The chips read, write and erase approximately 7 times faster than Flash memory, and also use less power. Sooner rather than later Samsung or the other PRAM producer Numonyx will put the chips in SSDs that can read and write at around 1GB per second.
From what I can see in the graphs the USB3 HDD is indeed faster than on USB2 because of the bandwith; the SATA HDD is about the same on SATA 2 and 3, but also pretty near USB3. The title is implying superiority of USB over SATA when clearly the HDD is the limiting factor.
It's relatively straightforward to add more parallel channels to an SSD drive and increase bandwidth. In the long run, there isn't even much of a cost difference to make the same capacity SSD drive fast enough to max out SATA 6. (the main cost driver of SSDs appears to be the cost of the flash chips themselves)
So bring on the new drives that can max out SATA 6! Right now, you can get comparable performance if you put two or four high end SSDs into a RAID 0 array. However, there's a lot of problems with doing this : you have to fuss with software drivers, certain SSD features aren't supported very well (like Trim), and there are bottlenecks in motherboard RAID chipsets because spinning disks were never this quick. Dedicated hardware RAID cards cost $300-$1000, making the cost rather steep for most users. Finally, while SSDs probably are inherently more reliable in the long run than hard disks, it's not a good idea to build a system that depends on 2-4 separate drives, a motherboard chipset, and potentially buggy drivers or else your data is hosed.
So I'm very much looking forward to upcoming SSDs like the Vertex 2 that should be able to max out a SATA 6 link. That is, once the SATA 6 motherboards become relatively common.
The big question with USB 3.0 is the price. That is the big advantage of USB over competitors like FireWire. Cables, host controllers, devices, hubs, everything is cheap. USB 3.0 looks a lot more complicated. The cables are much thicker with more wires and shielding. A USB 3.0 hub has to contain everything a USB 2.0 hub does, plus the new SuperSpeed part which is no longer just a dumb hub but more like a switch or router.
My issues with USB 2.0 are not so much about speed:
1- there's that ridiculous fudging about hi-speed, full-speed... is USB 3.0 **ALWAYS** USB 3.0, at last ?
2- I've got a bunch of 2.0 stuff (whichever 2.0 that was) that only works if I set my PC's USB ports as 1.0 only.
3- Even 2.0 stuff that kinda works has a way to make any non-intel-chipset PC freezy-jerky
4- I very rarely got anywhere near the supposed speed of 2.0 anyway.
In the end, I'd rather have a reliable, compatible, no PC freezes connection, than a "if everything works well" (read: rarely if ever) 10x faster one.
The Cloud - because you don't care if your apps and data are up in the air.
I don't know why the editors didn't include a link to it, but AnandTech has a much better review of the SATA 6G-equipped motherboard and its performance; one that actually gets around to doing real-world tests and not just synthetic tests. It turns out that the 6G Marvell controller is slower than the standard Intel ICH10 controller in virtually all cases. Until someone integrates SATA 6G in to a proper motherboard chipset, it's not just performance limited, it's performance degrading.
article title is misleading, it should be "usb 3 sucks, sata6 is amazing"
SURELY NOT!!!!!
It's faster. In their tests, they were getting 140MB/s transfers through USB3 to a single drive. I have two (older, slower) drives that can, between them, saturate a FireWire 800 bus giving me a total throughput of a shade under 100MB/s.[...] It is clearly superior to FireWire 800, which is the fastest FireWire you'll find on existing systems.
Well, it's 40% faster over the bus, but beyond that, is there anything inherit in the standard that's better?
Future ubiquity will certainly help, but if USB3's CPU overhead is as bad as USB2, it will mean the 3 out of 4 cores will be used by transferring at top speed.
Another possible advantage of FW is that it can provide a lot more bus power. USB can deliver 2.5W at 5V; FW can deliver 10 to 20W on average, but can hit 60W with a 30V rail. If you design your external device correctly, it means you don't need to connect a power supply when attached via FW.
1. My USB hard drives run consistently at ~25 Mb/sec. I have several types from different mfgrs and they all have the same transfer rate. I also have several 100 Mb Ethernet dongles from different mfgrs and they are all quite capable of saturating the network.
2. Not my experience at all! I segregate devices and hook them up to different hubs, but that's all.
3. Not my experience at all! I have nVidia chipset motherboards and NEC PCI cards that do USB just fine.
4.See #1
You don't mention anything about operating systems. I run Linux and USB works great for me.
The inevitable 10 Gbit Ethernet dongles will be limited by USB speed.
" ... When connected to a USB 3.0 port, the external hard drive was about 5 -- 6x faster versus connecting over USB 2.0, with total throughput in excess of 130MB/sec. On the other hand, benchmarks with Seagate's new Barracuda XT SATA 6G drive show little performance difference but a burst rate that is off the charts. ..."
So, the USB 3 will be attractive to consumers, with big, impressive numbers written large on boxes in stores everywhere, and the SATA 6G will be attractive to content creators (high end video production, etc). USB 3 will be cheap, and SATA 6G will be not-so-cheap.
About 99 out of 100 moderately clued in techies could have guessed the outcome of this one.
[Fudges around in toy box under desk ... pulls out crystal ball ... can barely discern "hippy type art school grad" reading AmandTech article dated Feb 2010 ...]
"Yeah, but wait ... it says here that if you load up the USB 3 with more than one device, they both really slow down, but my film lab's SATA 3G just keeps on truckin' when you daisy-chain them ..."
Yawn.
What, touchscreen slow? Seems like you never tried to transfer 3GB of photo's to a PC that only has a butterfly...
but USB3 is probably *slower* than eSATA when used with external SATA HDD.. and most of motherboards already have that connector.
Firewire 1600 / 3200 is better as it uses the same cables and ports as firewire 800. USB 3.0 needs new cables and ports also how high is the cpu load?
USB 3.0 provides more juice compared to 2.0. You could probably plug and power up a 3.5" desktop drive (assuming the enclosure has the circuitry to use power over USB) and it will run it without the need for brick adapters.
If you owned a firewire 800 disk drive, you would be smiling like me now.
When FW1600/3200 gets out of door, it will be same endless saga again since they will beat USB 3 too. They should also check the load on host CPU while doing those USB 3 speeds. Intel's standard is still host (CPU) controlled. Surprised a bit?
The speed of the USB connection will make NO difference in practice. Laptop machines have serious I/O bottle necks that typically don't allow the *sustained* I/O speed to exceed about 32 Megabytes per second. We started to call that the Galactic I/O Speed Barrier.
Excuse me, but please get off my Pennisetum Clandestinum, eh!
Sure you can yank up the bandwidth of USB3, but as long as you're stuck with PIO that isn't much of a gain. I would rather have USB3 have a DMA extension for really fast transfers instead of having to have the CPU wake up for every little I/O operation. On a related note, does anybody know any laptop brand that sells computers *with eSata*? That would be awesome.
Quantum hacker.
I wonder if even Intel's heart is in USB any more. USB 3 sounds considerably more complex than previous versions, not just for the chipsets but in terms of the cost of cabling etc. I wonder if the tech is going to see serious adoption. Intel are already talking up Light Peak which has a potential for insane transfer rates. I expect USB will be around for a long time yet, but I wonder if USB 3 will have time to become established before something much better appears.
Isn't this the exact same controller which was determined as bug-ridden and had to be removed?
- Faulty Marvell Chips Delay SATA 6G Launch
- SATA 6G launch delayed to do tech issues on nearly all P55 boards
- PC Perspective forum — SATA 6G delayed due to controller issues
- Marvell’s faulty 88SE9123 (SATA 6G) controller
Yup, that's the one. Buyer beware.
USB 3.0 does away with polling and introduces an interrupt-based transfer model, so CPU usage should no longer be an issue.
I ordered a new system based on an Intel CORE i5 750 2.66GHZ CPU running on the Asus Xtreme Design P7P55D-E Premium w/8 GB DDR3 1333 Mhz ram two days ago, and have been monitoring the net for signs of this mobo to actually hit the shelves. I will be running this with an unremarkable 64 GB Patriot SDD as the boot drive, until the new SATA 6 Gbps SSDs come out - which could take a awhile I imagine. I expect blazing speed from this platform, and can hardly wait for it. The only unknown is when will the mobo arrive. If it drags on and on, at least there is the option of an add on card that will convert one of the other ASUS X58 boards to USB 3 & SATA 6. I just hope I haven't made a mistake with the decision to wait. The P7P55D-E Premium motherboard will retail for $299 while the U3S6 add-on card will be $29.
Here are a host of links I collected on it this morning...
Asus Unveils USB 3.0 Motherboard
Asus Xtreme Design P7P55D-E Premium
The motherboard, unveiled Wednesday [October 28 2009], is 4.8 inches by 3 inches and is scheduled to be available next month for $299.
October 30th, 2009
USB 3.0 and SATA 6G Performance Preview - ASUS brings the goods
the P55-Express based P7P55D-E Premium is very close to hitting the market.
October 29th, 2009
USB 3.0 and SATA 6G Performance Preview
October 29th, 2009
This Is The First USB 3.0 Motherboard
October 28th, 2009
ASUS debuts USB 3.0 motherboard and add-on card
The P7P55D-E Premium motherboard will retail for $299 while the U3S6 add-on card will be $29. Both will be available November.
October 28th, 2009
ASUS brings the first mobo with SATA 3 and USB 3
October 28th, 2009
ASUS P7P55D-E Motherboard Offers USB 3.0 and SATA-III 6G Performance
North American Availability
The P7P55D-E Premium and U3S6 expansion cards will be available at ASUS authorized retailers early November at $299 and $29 respectively.
A 50MPH car isn't going any faster when put on a 100MPH highway.
There probably won't be FW1600/3200 at all. It'll be abandoned for USB.
Sure there may be a device or two created by some not so bright producers but the reality is they'll fade away and be forgotten.
It doesn't make a blind bit of difference how good something is, it only matters that average punter will buy it. USB will be everywhere, firewire will not. I stopped giving a shit about competing standards years ago.
Deleted
Like he said... price.
You can engineer whatever you want, but USB delivers 99% of what most people want at a fraction of 1394's price. Nobody buys a new pc so the processor can sit on it's ass turning watts into degrees.
There are plenty of use cases for Firewire, just not in the consumer space.
Platform advocacy is like choosing a favorite severely developmentally disabled child.
That's a pretty offensive sig. you have there, sport.
Can the fabric between these different IO endpoints be set by an application running on the CPU to move data between endpoints, say USB and SATA, or perhaps even network and SATA or USB, then get out of the loop? Configure the switch to move data between endpoint devices, without the CPU required to process the data at all until the transaction ends, or if an exception is thrown?
--
make install -not war
Having written an AHCI driver and worked endlessly on USB driver code there's no real point comparing the two. SATA is far, far, FAR more reliable. End of discussion. The USB chipset specs are horrid and the chipset implementations are even worse. Most chipsets barely pass through standard I/O operations properly and rarely deal with things like disk synchronization or even proper serial number reporting (for the USB bridge chips). USB has far higher cpu processing overheads and the DMA specs or so bad the driver often has to create bounce buffers. Command queueing overhead for a USB chipset is ridiculously huge compared to SATA chipsets.
USB is fine for a portable drive but only a complete fool uses USB if they need reliable mass storage.
E-SATA has its issues but they are nothing compared to the mess you get when you connect a drive up through USB. Frankly the only time one hits the 300 MByte/sec limit with today's SATA/E-SATA in a way that actually negatively impacts a production system is when one is talking to multiple targets over a port multiplier, on a single SATA port. The real need for 6GBit E-SATA is to better support port multipliers and not so much for SSDs.
While it is true that an SSD can hit the current 300 MByte/sec limit over SATA, there aren't really any realistic production loads against single drives (verses port multipliers) where an improvement in the bandwidth would actually improve the machine's ability to handle its workload.
-Matt
FW1600/3200 will be good for couple of purposes, not for ordinary users cheap Taiwan "backup" disks or portable disks. It was always the case for FW400/800 too, it is just rich Apple owners wanting better quality with lower (or non existent) CPU overhead opted in for Firewire or dual (USB2/FW) drives.
Firewire is a must for TV stations, audio engineering products, professional photography, in field workers. Obviously, while there is need, especially Sony PROFESSIONAL (gotta use CAPS on /.) will ship the products. They won't put all their money to FW1600/3200, they of course know the issues with Firewire and why it won't replace Joe Sixpack's USB2 (3) drive. I use FW800 drive for actual data, 2 of them in software raid configuration, for SD editing but my "time machine" backup drive is in fact, set of 2 USB2 drives with software RAID mirroring.
What you must understand is, it doesn't make producers "less bright". It is a different World with very different (sometimes crazy) needs. You wouldn't need 300MB (yes, MB) per second stable bandwidth for data but a 2K/4K movie editor can't do his job without it. So, there comes SCSI and some idiot out there will sure compare it to SATA2/3.