64GB Survivor GTR poor performance

[Tony_UK]

I recently bought a 64GB Survivor GTR. Currently I'm using it alongside my 8GB Survivor GT.

 

I usually use Laplink Gold to copy changed files from selected folders on my PC to both these USB drives - though not at the same time. The transfer rate for the 8GB GT is around 22Mbps. The 64GB GTR is very slow to get going and 'stuttery' when being written to. The transfer rate is lucky to get up to 11Mbps.

 

Is this normal?

 

I guess I've been spoilt by the GT 8GB performance and, I presume, the GTR is fast reading but nothing like as fast when writing.

 

If it's helpful, I downloaded ATTO Disk Benchmark and ran it on the 2 drives. The results are GT 8GB and GTR 64GB. Are these the sort of expected performance figures.

[RAM GUY]

Tony

Can you test the GTR drive on another system just to test?

And when its connected are you installing the drives side by side or in different ports?

[Tony_UK]

How very interesting that little exercise was!

 

I run the Survivors using USB extension cables to avoid messing around at the back of the PC so I did two more benchmarks with the Survivors swapped over in the extension cables.

 

I also have a Samsung NP-N140 Netbook bought last Happy Holidays!, running XP Home. I connected in turn the Survivors to that PC direct into the USB port, one at a time and ran the benchmarks. These showed an increase in read/write performance for the GT 8GB of around 50% & 150% respectively. For the GTR 64GB the read/write performance improvement is around 100% in both cases. That is quite an improvement!

 

There seemed to be a number of possibilities, the extension cable, the USB card, the driver.

 

So I ran some more benchmarks. Firstly connecting the Survivors directly into the USB port at the rear of the PC. Then connecting the Survivors directly to a USB port at the front of the PC.

 

After this I moved the extension cables to two USB ports that are, as I understand it, direct to the motherboard rather than via a PCI extension card, and ran two more benchmarks.

 

All the benchmark results are summarised in the zipped Excel file attached. It clearly shows moving the extension cables to the motherboard connected USB ports has improved write performance by 100% or more, read performance around 75%. This is not as much as on the netbook, however that may well be the affect of not using the cables on the netbook.

 

Having seen this sort of performance improvement I tried Laplink again. Oh dear!

 

Copying 77 files totalling 503 MBytes took 20m 43s, at a transfer rate of 3,400,300 bps! I then tried copying the same 77 files using Windows Explorer copy and paste - 28 seconds! So Laplink is some 40 times slower than Windows Explorer, clearly a total nonsense. What on earth it's doing to give such poor performance I have no idea.

 

Many thanks for pointing me in this direction. I must admit though it still doesn't seem to explain why using Laplink the GTR was much slower than the GT. I'd guess something to do with transfer sizes or the like.

 

Clearly I need to find some more efficient multi-folder synchronization software.

 

I rather enjoyed the exercise :biggrin:

Corsair Survivor Performance.zip

[RAM GUY]

Thank you for sharing the information, but have you talked to Laplink to see what they say?

[Tony_UK]

I've put a ticket in with Laplink.

 

A quick test with Laplink and the two Survivors shows those same 77 files, which take over 20 minutes on the GTR 64GB, take 2min 10sec on the GT 8GB, but Windows Explorer (26 secs on the GTR) takes 31secs on the GT.

 

All very odd. :confused:

[RAM GUY]

Sounds as they may need to update their software to work better with newer flash controllers, but please let us know what they say. Unless there is some sort of CRC Check being done in the copy process is about the only thing I can think of that may slow it down like that.

[Tony_UK]

To keep you in the picture.

 

Laplink asked for more Windows Explorer (WE) and Laplink (LL) tests with all Antivirus, Firewalls, and 'unnecessary' services disabled/not running. I did this (with the Internet disconnected!), it was only 3-5% faster, i.e. no significant difference. I then put my PC back to 'normal' and re-ran the tests, as expected same results as originally.

 

Although they hadn't asked I also did WE and LL transfers to my 3 external drives of exactly the same files. LL was about 10% slower than WE. Not unreasonable I thought. But nothing like the 320% (8Gb) and >3,900% (64Gb) slower with USB drives. As I said to Laplink it seems their handling of USB drives seems somewhat poor, even more so with larger ones.

 

I also commented to Laplink that when writing to the USB drives the CPU utilisation is c.5-10% (8Gb), 0-5% (64Gb) suggesting I/O is the delaying factor, certainly not processor.

 

Anyway, they responded a few days ago to say that their support have taken in all the data and passed it to their QA department to take a look at.

 

I await the outcome with interest.

[RAM GUY]

No problem and thank you for the update and please if you do not mind let us know what they say or do to address this.

[Tony_UK]

Not having heard anything from Laplink I asked them yesterday if there was any update. Their reply was "No new developments have occur for this product. If any happen you would see an email from our system letting you know." That can be read as saying nothing is happening and quite possibly nothing will happen.

 

I hope I'm misreading between the lines, but...

[RAM GUY]

Tony

I do not expect them to change something like that with out some serious testing, plus is an issue with one device so I would not expect it to happen quickly.