Build your own External Hard Drive

[Cr00zng]

Thanks Essenbe...

That difference in time seems negligible on the surface, but...

If my calculation correct, there's about 10% throughput difference between the drives if you divide the size of the image with the time in seconds using your numbers.

Internal:

8.1 / 175 = 52 MBs

External:
8.1 / 176 ~ 46 MBs (rounded down from 46,022,727.27272727 bytes)

Macrium collects the data, compresses the file, and sends it to the drive; the chances are that it'll not max out the internal or the external drives' throughput. In my opinion a benchmark tool would be more indicative of the performance differences than a real-time test...

 

[essenbe]

Cr00zng, see if this helps but I still say Samsung's are the only drives I remember that write faster than they read.
Internal Samsung


External

 

[Cr00zng]

That's some interesting performance results with your Samsung drives...

Upto 8K size the internal drive is between 10-30% faster than the external one. Beyond the 8K mark, the advantage is flipped in favor of the external drive and it is between 1-3% faster. This might be due to ATTO that I haven't been using.

I didn't see any drive that would write faster than read, odd to say the least...

I'll retest my Seagate drives with ATTO and compare the results with yours.

 

[essenbe]

That's some interesting performance results with your Samsung drives...

Upto 8K size the internal drive is between 10-30% faster than the external one. Beyond the 8K mark, the advantage is flipped in favor of the external drive and it is between 1-3% faster. This might be due to ATTO that I haven't been using.

I didn't see any drive that would write faster than read, odd to say the least...

I'll retest my Seagate drives with ATTO and compare the results with yours.

Here's my Seagate 500GB Barracuda 7200.12

 

[Cr00zng]

Well, I've done my testing as well and I am not sure just how accurate ATTO is?

Internal Seagate 1 TB 7200.12:



External Seagate 1 TB 7200.12:



The external drive is faster than the internal one, up to 2 KB between 200-300% faster, which is the opposite of the CDM results. With the exception of the 64 KB, all others are faster with the external drive, albeit by a much smaller margin.

On this system (Intel P67) the only difference is the data bus where the drives connected to; the internal one is hooked up to the SATA 2 port, while the external one connected to the eSATA 3 port provided by the Marvell 91xx chipset. You may have a similar setup with your Asrock P67 board...

It's possible that the Intel SATA 2 port still have some minor performance issues, I don't know and it is just a wild guess...

 

[essenbe]

I have the same setup. But both of mine are hooke up to the P67 sataII ports. The external is through a front panel connector on my case.

 

[essenbe]

See if this is better. But the internal is 6 mos. older and has a lot of data on it. The external is new.

Internal


External

 

[Cr00zng]

I'll test the drives tonight; mine had been purchased at the same time and presumably, they are the same.

I didn't even tell you, but you have already convinced me that next time I should buy Samsung SpinPoint drives...

 

[essenbe]

I'll test the drives tonight; mine had been purchased at the same time and presumably, they are the same.

I didn't even tell you, but you have already convinced me that next time I should buy Samsung SpinPoint drives...

They've been good to me, so far.

 

[pparks1]

Thanks Essenbe...

That difference in time seems negligible on the surface, but...

If my calculation correct, there's about 10% throughput difference between the drives if you divide the size of the image with the time in seconds using your numbers.

Internal:

8.1 / 175 = 52 MBs

External:
8.1 / 176 ~ 46 MBs (rounded down from 46,022,727.27272727 bytes)

Actually, your calculations are not correct

8.1 / 175 = 0.04628571, ...which is about 46.2MB/sec

8.1/176 = 0.4602272, which is 46.02MB/sec

So, actually a very small difference...or about 1 second over nearly 3 minutes.

 

[Cr00zng]

Oops, thanks for the correction pparks...

The HD Tune benchmark results are confusing for me.

Seagate 1 TB internal on a SATA 2 port:



Segate 1 TB external on eSATA 3 port (Marvell 91xx):



The identical drives had been purchased at the same time; the internal drive has about 4% usage while the external does 38%. Yes, I did rerun the tests and had the same results.

Why would there be such a wide gap in access time and burst, even if they are connected to different SATA ports?

Just for curiosity, I've tested my older machine (specs in my profile).

Internal Seagate 320 GBs on SATA 2 port, the drive is four years old:



And this is the external Seagate 1 TB (different model# from the above) on eSATA 2 port, the drive is about 7 month old:



I am puzzled... the four years old drive has better access time than the 2-7 month old?

 

[pparks1]

The thing with benchmarks is that they are not always consistent.

For example, run HDTune 10 times in a row. Your numbers won't be the same...but they really should be.

When I build my current machine, 2 years ago, I went with a 1TB Seagate 7200.12. It performed great, but benchmarks it had about 25ms access time for some reason. No matter what port, what cable, what drives, or what OS.... (vista or 7), I showed this same slow access time. Real world usage didn't seem to indicate anything was wrong. It was this very problem that brought me to this forum.

In the end, I ended up buying a WD Caviar Black 1TB for my new machine, and it was about 15Mb/s slower transfer, but was around 8-10ms access time. I put the Seagate 7200.12 in my older Linux desktop, and it came back with 12ms access times and has performed just fine.

Benchmarks for me, are not the be-all end all for my performance. They are simply an indicator showing that nothing is terribly wrong or broken.

If you put that same 2-7 month old hard drive in 5 different boxes, you would likely get 5-7 different results. Don't fret over 1-2ms, it's simply just not worth it.

 

[essenbe]

I have a very similar opinion of benchmarks.I started off by cheking my SSD. With AS SSD the numbers were all over the place. I'v seen scores of 335 and other times I've seen scores of 440. Yet, in actual useage I could tell no difference. Startup times were the same and apps opened the same. I've come to the conclusion that benchmarks have no bearing on actual usage.

 

[James Colbert]

Thanks for the heads up on the Samsung/Newegg special. I picked on up just before the offer expired. I just now initialized and created a partition on it to make sure it was operable (life is busy...). It is. Great deal. Thanks!

 

[linnemeyerhere]

For the heavy data whores and you know who you are this bad boy or the like is the direction I would tend to go. Smallish foot print with massive expand ability.
Sans Digital TR5UTP Hard Drive Enclosure - 5 Bay, 3.5, SATA to USB 3.0eSATA, RAID, 2-Port eSATA PCIe Card Included at TigerDirect
4 bay in black.....nice!
http://biz.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=6610223&CatId=2780

 

[essenbe]

For the heavy data whores and you know who you are this bad boy or the like is the direction I would tend to go. Smallish foot print with massive expand ability.
Sans Digital TR5UTP Hard Drive Enclosure - 5 Bay, 3.5, SATA to USB 3.0eSATA, RAID, 2-Port eSATA PCIe Card Included at TigerDirect
4 bay in black.....nice!
Sans Digital TR4M-BP Hard Drive Enclosure - 4-Bay 3.5 SATA to eSATA, RAID 0, 1, 5, 10, JBOD, SATA 6G PCIe Card at TigerDirect

I've thought about it but don't know If I want to go that route. But I do have enough hard drives to fill them up.

 

[Cr00zng]

When I build my current machine, 2 years ago, I went with a 1TB Seagate 7200.12. It performed great, but benchmarks it had about 25ms access time for some reason. No matter what port, what cable, what drives, or what OS.... (vista or 7), I showed this same slow access time. Real world usage didn't seem to indicate anything was wrong. It was this very problem that brought me to this forum.

Maybe I should've listen and leave the machine alone, but no...

I moved the internal drive to the Marvell based SATA 3 port and might as well, changed the cable:



Other than the huge increase in burst rate, which is doubtful at best, there's no change in the benchmark performance. Or for that matter, in the real world usage performance as you've stated.

Benchmarks for me, are not the be-all end all for my performance. They are simply an indicator showing that nothing is terribly wrong or broken.

If you put that same 2-7 month old hard drive in 5 different boxes, you would likely get 5-7 different results. Don't fret over 1-2ms, it's simply just not worth it.

1-2ms agreed, but when the difference is about 6ms that does indicate some issues; however, the difference is still not to the point to be worth for looking into it further. Especially when the drive works just fine otherwise.

The issue could still be the benchmarking tools and/or the drive quality in itself. I could swap the drive to verify it, but at this point it isn't worth for the effort...

 

[Hopalong X]

FYI

Newegg U.S.- 20% off external enclosures or docks. June 27-28.
Not sure about Canada.

I also posted on the Deals thread in Chill Room.
It seemed appropriate here also this one time.

Newegg.com - External Enclosures, External Hard Drive Enclosures
--

 

[pparks1]

Other than the huge increase in burst rate, which is doubtful at best,

Yeah, that burst rate is obviously not realistic...clearly the benchmark is not working properly or measuring remotely accurately.

1-2ms agreed, but when the difference is about 6ms that does indicate some issues; however, the difference is still not to the point to be worth for looking into it further. Especially when the drive works just fine otherwise.

Exactly, my random access times on my Seagate 7200.12 read in the 20-25ms range on my box. I drove myself nuts and replaced it with a WD Caviar Black. All-in-all, I can tell no performance difference between the drives on the actual machine. That Seagate in another machine is fine. So, I'm unsure if there were any actual issues, but I felt better seeing that my access times were lower with the WD. And just being a storage drive in the end (as I'm now using an SSD for the OS)...the access times are far less important now.

 

[essenbe]

Newegg Samsung Spinpoint $55.00