PCI Ultra ATA Controller verse USB 2.0

PCI ultra ATA 133 controller card from PROMISE say the data transfer rate is 266MB/sec while USB 2.0 transfer is 480MB/sec. Does this means external HDD using USB 2.0 is faster ?

hhhengAsked:
Who is Participating?

[Webinar] Streamline your web hosting managementRegister Today

x
 
Kyle SchroederConnect With a Mentor Endpoint EngineerCommented:
Theoretically, yes...but in real life, probably not.  The 400 MB/sec is not technically correct, USB 2.0 has a (theoretical) maximum bandwidth of 400 Mega*BITS* (Mb, note small "b") per second, so its really only 50 MB/sec (8 bits in a Byte).  In real application, this 400Mb number is likely not possible to achieve.  Whats more important is the RPM factor of the drive; a 5400RPM internal drive *might* be beaten by a 7200RPM external USB2 drive.  There are other factors as well, including cache RAM size on the drive, GB/platter (data density), etc.

-dog*
0
 
jhanceCommented:
When drive manufacturers make "real" USB2.0 drives, then this will be possible.  But today, a USB external drive is just a plain old IDE drive in a USB enclosure.  

So regardless of how fast the USB interface is, the whole things CANNOT work faster than the speed of the drive in the box, which is likely to be ATA/100 but certainly no better than ATA/133.

Eventually, IDE will be dropped in favor of native USB interfaces on internal drives but I'm not aware of any on the market today.
0
 
magarityCommented:
USB 2.0 transfers up to 420Mb/sec.  Just like a network transferring at 100Mb/sec, note the lowercase 'b', as in bits, as opposed to 266MB/sec, capital 'B' for bytes, for the ATA133.  If we divide 420 megabits by 8 to get bytes, we get a theoretical maximum of 52 megabytes.  However, this is throughput, not real world.

Discourse on minutiae:  USB is 'polled', so that the controller asks each device in turn it there is something to send (or demands the device receive).  When the device is transferring, the transfer time is a very limited length, as I recall it's 10 milliseconds.  Then the controller interrupts the transfer to go around and ask all the others if they need to send something.  If they do, they take their turns.  Even if there are no other devices, the controller has to ask the channel if there are any new devices it doesn't know about.  So the transfer is constantly being interrupted.  This is all opposed to an IDE/ATAPI device which can lock the bus and spam data as hard and fast as it can.  Keep in mind that the 266 figure quoted by the manufacturer is also a theoretical number and won't happen in the real world.

Anyway, just a couple of weeks ago, I got a new USB2.0 controller card and a couple of USB2.0 drive enclosures to test out.  The Fujitsu 20GB laptop sized drives transfer at 7.8MB/sec, sustained, according to SiSoft Sandra's disk benchmark test.  On USB1.1, the same drives transfer at 700KB/sec.  7.8 is not at all bad and perfectly usable as a secondary drive to a laptop or even backup storage for a desktop.  In these days of cheap 30MB/sec ATA100, 7.8 may sound sucky, but it depends on where and doing what.

regards,
magarity
0
Will You Be GDPR Compliant by 5/28/2018?

GDPR? That's a regulation for the European Union. But, if you collect data from customers or employees within the EU, then you need to know about GDPR and make sure your organization is compliant by May 2018. Check out our preparation checklist to make sure you're on track today!

 
hhhengAuthor Commented:
So USB2 drive is actual transferring data at about 52 megabytes/second. As compare to PCI controller card at 266 megabytes/second. This a lot of difference between 52MB and 266MB despite the fact that both are using PCI cards. The difference, I reckon that PCI controller card uses a 80 wire cable which is better than USB2 4 wire cable.

So the right answer is PCI connected HDD is faster than USB2.0 external HDD, if considering all factors are the same ( cache RAM, rpm,etc )

I've lost some screws in my head and now afraid to award points. What interest me most is reading all the comments which provide me with lots of informations/knowledge with intangible value. Like "magarity" despite labelling me as a jerk has given a very lenghty techno-savvy comment. The mild and meek "jhance" who was angered by me for giving points to a "copy-cat" answer. And "dogztar" whom I made uncalled remarks has come out to help me again.

Can some EE moderator help out here in awarding the points.

Regards & cheers..
Heng



0
 
Kyle SchroederEndpoint EngineerCommented:
jhance, mild and meek?!?!? HAHAHAHAHAHAHAHAHA

Heng...we are helping because that's what this place is about..its also a good opportunity to display technical knowledge amongst peers.  So we have let that other stuff go and are just trying to help (which is after all what we're doing here...)

But please, try to keep the personal comments to a minimum, or not at all, OK?

-dogztar
0
 
hhhengAuthor Commented:
"mild and meek" was first crooned by a "rose among the thorns" and now I realised this is true.

Hope a little itchy and bitchy comment didn't hurt that much. "David Letterman" does that all the time and the American audience loves that guy. HE HE HE HE HE

Cheers..
Heng
0
All Courses

From novice to tech pro — start learning today.