Is this normal? SSD/USB3.0

Currently I have a Patriot Inferno SSD in my N53
and the original WD 7200RPM HDD in an external USB 3.0 enclosure

I was getting like 10MB/s transfers from SSD to HDD on a 2.0 enclosure
so I bought myself a SilverStone 3.0 enclosure and got the following results
Transferring from SSD to HDD = around 20MB/s, first second = 120MB/s
Transferring from HDD to SSD = around 60MB/s, first second = 400MB/s
The funny thing is at the beginning for a second it transfers at super fast speeds then slowly drops to 20 and 60MB/s

Is there something limiting my speeds or is this how it's supposed to be?
My USB3.0 driver is up-to-date
(Yes, it is plugged into the 3.0 port)

 

Those look low (like with my Inferno, btw - but no USB3 - eSATA), but normal, I guess.

Depends on what type of files you're moving - a single large file (and the type of file will make a differernce too - tiff, zip, rar, iso, etc.), medium size files (documents, mp3's, pdf's, etc.) and/or a combination like program files (large zip, exe, dll & small exe, dll, txt, and ini files).

Can you try your exact same drive (and data) in an eSATA enclosure instead? I'm sure it should be faster (especially if you're moving a lot of comparatively smaller files).

Good luck.

 

Hey. Great little Asus you got there. :^)

I've never seen more than 66M/sec on USB 3.0 bus, even mirroring SSD to SSD. Both are Sandforce (OCZ Vertex2) and the SATA connected one is running AHCI mode whose reads hover around @ 266M/sec.

USB 3.0 is not 10x faster. Not even 4x faster, as you're finding out. It's not an eSATA killer by any stretch, and if you're going to leave external HDDs hooked up most of the time, I would also suggest getting a eSATA chassis or two. (You can get a multi-HDD chassis too, the N53's eSATA chip supports Port Multiplier.)

I had high hopes for USB 3.0, but as time goes by I'm wondering how many people will ever support it before something like Light Peak comes out. Remember that USB was an Intel creation, just like Light Peak.

The speed peaking when you first start that's actually two things: data caching, and the way rounding maths work. Same reason sometimes you see initial transfers on a download grossly exceed your bandwidth, then level out.

 

Now you have 90MB/s on USB 3.0 - with transfers chart

see also eSATA versus USB 3.0 and the extremetech article there .. Looks to me that if your enclosure fully utilizes the bandwidth, USB 3.0 doesnt even have to be twice as fast as Usb 2.0 to be better than eSATA.

 

I got nearly 130mb(OCZ Enyo SSD) write speed with NEC USB 3.0 chip. I think it depend on USB chipset and PCI-E bus speed your chip connected.

 

I wouldn't hold my breath on LP taking over anytime soon. But as for the increase in speed from USB 2.0 it's largely theoretical at the moment, but it will increase in time.

 

The first link is 90 MB/s reads (not writes).

The second link states that with eSATA, writes are faster, while reads are slower compared to USB3 and similar/identical HDD's.

The reason I use external HDD's is for writing to them (they're Back Up devices) - eSATA is still untouchable. And, as someone already mentioned, with SATA3 coming soon - they will remain untouchable too.

Until LP, of course.

 

Good points .. the article I linked didn't provide a good write-specific comparison. YouTube - USB 3.0 Speed Test vs USB 2.0 eSATA Internal connection (NCIX Tech Tips #71) -- here esata and usb3.0 are running neck-and-neck except with batches of small files, -- which is useful for some (but not all) back-up mechanisms.

eSATA might be a bit ahead of the game but not terribly (very nearly equal/slightly worse than USB 3.0 performance on large file transfers, and queued/batch file transfers its better but thats a rarer use) -- and while eSATA may someday pickup the SATA 600 signal, I've seen no news about that on the horizon. Also note that the problem is clearly not related to the available bandwidth in the interface, but in the USB controllers -- otherwise the read speed would not be higher than the write speed on the first link I posted.

Eventually USB3.0 shouild always be slightly faster than esata 3Gbps, and these chips are only utilizing 40% of the increased bandwidth available since usb 2.0 -- so it really could end up 5% faster than esata 6Gbps, once they start making those.

To be fair, right now you can always just rig your internal sata3 ports to the external face on desktop PCs and trounce USB 3.0.

 

I think eSATA and USB 3.0 serves different purpose. The advantage of eSATA is that NCQ should work as it seems to be just a different form of cabling but the protocol is the same as SATA. I don't know how USB 3.0 can support this. NCQ is important for multitasking scenario(small file, random access etc.).

Though at the moment, eSATA is still being used as kind of tape backup rather than 'just another disk' outside the chasis.

 

Are we forgetting something here folks? USB 3.0 is an all-in-one port! Without it, all the aforementioned jack rabbits ports are just treading water.

 

Who ever tells people this is ridiculous. USB 3.0 is ~4.8Gbps(or whatever it is). It runs at its full speed all the time unless you get an express card. That will force it to 2.5Gbps because of the PCI-e 2.5Gbps bus speed. Second eSATA is always 3Gbps unless in a pci-e bus. Than it runs at the 2.5Gbps. Your experiencing the hardrive slowness. I had to go into the wikipedia to update it because some tard said eSATA is limited to 119MBps, which is garbage.


Also what are you talking about. usb 3.0 doesn't even have to be twice as fast as usb 2.0 is be better than eSATA?????? USB 2.0 is 30MBps....eSATA is ~300MBps that's 10 times faster than USB 2.0. Anyways i checked wiki again and the tard changed it back so off i go fixing the wiki because eSATA is not limited to 119MBps....morrons


http://en.wikipedia.org/wiki/Serial_ATA#eSATA
updated wikipedia again....

 

I hear you, it's a little funny.

well, USB 3.0 standard is at 625MB/s .. which is 8x6 = 4800ish Gbps. All communications standards are optimistic, even in the best cases ( things like internal SATA, fiber optic international cables, etc) you never expect more than 80% in practice.

With USB, the ceiling has been particularly overoptimistic and a 60% bandwidth is a more common number... so much so that you can read it in the wikipedia article on USB, under Traffic Speeds in Practice.

But, when any technology is brand new, the controllers can be sub-optimistic .. first generation almost always is .. so it really wouldnt have been any surprise at all for USB 3.0 to only show (perhaps) 40% of its advertised bandwidth. That's 250MB/s .. which is only twice a (mature) USB 2.0 controller's optimistic transfer rate. Later 3.0 controllers should rise in reported bandwidth, and probably plateau at 4x optimistic USB 2.0 (500MB/s)

--

With mass storage, a new problem arrises, the device controller must translate from the sata connector on the drive to the usb 3.0 -- especially at first, corners are cut on die, and transfer speeds drop .. asymmetrically, as can be seen by the fact that read and write speeds are so vastly different in the data above.

 

interesting setup here -- A USB key flashed at 371 MB/s! - BeHardware with specific usb 3.0 controller and ssd disk:

note the sequential writes running at 190MB/s over USB 3.0 -- see guys, it is the disk controller on the usb device, not the usb 3.0 bus itself!