USB Bottlenecking HDD Speed Question

If connecting an external HDD by USB 2, does it matter what RPM drive you get? For instance, won't a 7200 RPM drive get the same read/write speeds over USB as a 4200 RPM drive?

 

It's true that USB will restrict your sustained transfer speeds, to usually no more than 30-32 MB/s, but it's still worthwhile going for the faster drive for it's faster seek times.

 

My 7200 rpm 3.5" maxtor drive has double the transfer rate of my 4200 rpm 2.5" drive.

 

Are both over USB or what?

 

Even at therotical speeds you will not notice a diff ... 480 Mbps for usb 2.0 and even ata 66 is >500 Mbps.
EDIT: Even the WD800VE which is one of the slowest current notebook drives is much faster than usb.

If they are close in price go for the faster (IF you may take it out of the enclosure later), otherwise grab the cheaper drive

You are not going to see a difference ... seek times 'might' make a little bit of difference. The buffer will compensate for much of the seek time if transfering multiple files so even if you transfer lots of files at once(large database backups) chances are you will never notice the difference. If you use the drive as a media or file drive, never notice.

 

"Are both over USB or what?"

Yes - well, my 3.5" drive can use either USB or firewire, I generally use firewire but it clocks only slightly slower on USB (35 MB/s vs 40 MB/s). My 4200 drive clocks at about 21 Mb/s.

If you're backing up the computer, you will definitly notice a difference - as I said, there are considerable differences in the read/write times for my two external drives, which is definitely noticeable in rea life usage. If it is simply to watch movies or store small files, I wouldn't worry too much.

If your intent is to have something that remains in one place (i.e. at your home desk), just buy a 3.5" external drive combo - it's way cheaper than the 2.5" of comparible size and will perform much better too. If you need something more portable, the 2.5" will be a better choice.


"Even at therotical speeds you will not notice a diff ... 480 Mbps for usb 2.0 and even ata 66 is >500 Mbps.
EDIT: Even the WD800VE which is one of the slowest current notebook drives is much faster than usb."

HOwever there still remains a difference when using a faster HD externally despite this. This depends on several factors, including the drive, the chipset in the external enclosure, and buffering.

 

Well Cheffy your 4200 must be in really useless enclosure to get that low transfer rates.
I am sure that if you connected your 4200 drive to your 3.5" case you would see that there is no difference between your drives.

 

I'm sure that you have miscalculated.
The fastest HDD at the moment uses SATA-II which is 300MBPS,but the USB is 480MBPS.It's true that USB's protocol never let's the actual data transfer speed to reach 480MBPS and that USB is slower than 1394 (at 400MBPS) at transferring data.But still, the main bottleneck is the HDD itself and not the USB.

 

is this a 3.5" or 2.5" drive?
does it matter?
how do I find the transfer rates, ect.?

 

SATA-II is 300 MB/s (MEGABYTE/SECOND)
USB 480Mbit/s (MEGABIT/SECOND) thats 60MB/s (megabyte/s) max.

 

You are wrong, you are confusing Mega Bits (USB) with Mega Bytes (SATA)

 

Oh I see

 

Possibly, but considering my 4200 drive clocked only slightly faster when in my actual notebook (~27 Mb/s avg) I doubt it. That's inline with most other results I've seen.

 

21 --> 27MB/s is not a small difference

 

yeah people always confuse MB and Mb not to mention KB and Kb.
I just wish my 2567 kbps connection could fill my hard drive faster!

 

I'd have to double check, might not have been quite that high actually. But my point is that the USB connection will slow down a HD regardless of it's actual potential output. If I put my current 7200 2.5" drive into a top-end case, it will still function slower through USB than it does in my notebook, even if the potential throughput of USB is higher than my HDD actually transfers currently in the notebook itself!

 

That was my point exactly, any drive that is built in the last 5 years (ata66 or greater) will max out a usb connection. USB is a bottleneck to drives so it doesn't matter which one you buy for an external usb enclosure UNLESS you are doing backups of 10's of thousands of small files in which case the seek time MIGHT become noticable.
This is also why I recommend going for a faster drive ONLY if you might be taking it out of the case and using it inside a computer.

EDIT Sorry Muj, MBps and Mbps are easily confused. I shoulda written it out

 

Except for the fact that the transfer rate on my 7200 drive is still nearly double what it is for my 4200 drive - for occasional file transfers this isn't a big deal, but if I'm making frequent backups or transfering huge files regularily, it cuts the time in half, regardless of USB limitations. If you want to view pictures as thumbnails this makes a big difference too, they take much less time to load. I agree for most uses though there will be few issues. But as I said, if you do not need the portability of a 2.5" drive, there is no point in spending twice as much or more per GB of storage by buying a 4200 2.5" drive instead of a fast 7200 3.5" drive, especially when you will definitely be able to transfer much quickly with the 3.5".

 

The arguement is that when connected externaly even a 4200 2.5" drive has a transfer rate that is over 25MB/s (typical USB transfer rate) so the only difference you will see is any difference between access times.
For big files you shouldn;t be seeing any difference even if you are using a raptor since your USB connection will be the bottleneck.

 

No, not true at all. I personally have witnessed that the transfer rates are notably different, and this is hardly controversial, it is well established. Also, you'll barely get 25 mb/sec with a 100 gb 4200 drive. PErhaps it the new 200 GB 4200 drives will be better. A simple google of this will confirm what I'm saying by numerous sources on the net.

The simple fact is that it is not simply a matter hitting a wall with USB and no more performance is to be had. The way in which the data is transfered through USB always results in a decline in performance when compared to PATA or SATA due to rerouting through less efficient hardware. It will hit a wall eventually, but that is much higher than most notebooks are able to transfer INTERNALLY nevermind externally (~60 MB/s). In my experience there is generally about a 25-40% decline in performance through USB compared to internal bus transfer. This is why my 4200 drive transfers at 25 mb/s internally, and 20 mb/s externally, and my 3.5" drive would transfer 60 mb/s internally (in a desktop), and transfers 36 mb/s externally. Access time is obviously different, but I'm speaking about transfer rates, and not just with small files. Specifically larger files.

You can disagree all you want, but the numbers are there to show this is the case.

 

I would like to see your results for 4200 disks.

 

Unfortunately I don't think I kept a screenshot of my 4200 while it was still in my notebook. I can produce one for it externally at least, although I'll have to do it later as my HD is currently somewhere I won't be for about a week or so. If I'm feeling really bored and adventurous I'll swap it back into my notebook and test it, but don't count on it.

 

I don't notice any signifigant differance between my 5400 RPM Seagates (here they have the 120GB on sale, I just bought 1 of these) and my 7200 RPM Lacie. I always atributed the bottleneck to the USB 2.0 transfer speeds, but it appears that I may be mistaken.

I do have to say, my dv6000t with the t5600 and 2GB RAM with Vista and a 5400 RPM drive is faster (both seek and load) then my friend's Vaio, with the same specs except he is using XP and a 7200 drive (both Toshiba drives by the way). So I really wonder is the differance between 5400s and 7200 really worth the price?

I love my Seagates (I have 2 40 GB drives, and now the one 120GB drive on it's way) with the 5400 drive. I wouldn't hesitate at that price.

 

A check of benchmarks such as at Tom's hardware will show that there are some 5400rpm drives that consistently out-perform some 7200rpm drives.

rpm is only one contributing factor to hard disk speed. It's like measuring all cpus by clock speed alone. Other factors to consider are density, allignment, manufacturer (e.g. Seagate pwn Fujitsu) etc.

 

Not quite true, even the latest 160 GB 5400 drives are just lick below the last generation 7200 drives for nearly all purposes. On very large transfers there is some advantage for these 5400 drives, but the access time is still notably lower and for most purposes still slightly slower. Of course, the latest 7200 drives have now changed that.

 

There is a graph at the top of the page. Access time is clearly rpm related so no chance a 5400 will outperform a 7200 drive there. As for the transfer rates again 7200 drives have higher transfer rates in most cases with the occasional larger 5400 drive outperforming some older 7200 drives by a small margin.

 

about the la cie to seagate comparison stated above. Not really apples to apples, La Cie has let quality fall a lot. Check New Eggs user reviews if you need confirmation. Their specs have slid downhill too. Look at them on Toms charts.

Not sure if I hve the same models pegged on the charts but the seagate scores in the high middle for most metrics. La cie near bottom

 

I agree to a certain extent. But you need to bear in mind that the Lacie Portables are, in most cases, 5400 RPM drives. While they sit lower in the stack, if you actually look at their speeds and performances, the difference is quite nominal.

And I would never expect a Lacie 5400 RPM drive to outperform a Seagate 5400 RPM drive.

I would expect a Lacie 7200RPM drive to outperform my Seagate 5400s. And if you look at the charts, the Big Disc Extreme (Lacie) consistently outperforms the Seagate, albeit by a smaller margin then expected when you consider drive speed and the drive size (the Dig Disc is a 3.5, and the Seagate is a 2.5).