Did a quick set of tests today between my WD Green 2TB and new WD Black 2TB.

 

According to the reviews the random IOPS of the Black is roughly double that of the green, so I was expecting a healthy increase up to around double the performance.

 

Guess what? I got 195% the performance!

 

The test was writing a 1GB file to two shares from two machines. (So one file to disk2 and one to Media (Disks 3 and 4). The average of the write performance was taken both times (4 runs in total from two machines)

 

Green: ~6.7MB/sec

Black: ~13.2MB/sec (197%!)

 

Pretty large improvement for overlapped IO!

 

Well worth it IMO.

Something seems strange here. Both tests showed speeds of <15 MB/sec which is actually pretty poor. I have all green drives in my server and I typically see write speeds in the 25MB/sec range. The only way you would see an increase in write speeds is if the data disk AND the parity disk were both 7200 rpm drives. Black drives may very well be faster than Green drives but I suspect that there is something else going on here.

This test was with the WD drives as the parity drive?

 

What was the data drive?

 

What tool did you use the make the benchmarks with?

 

Also depending on RAM 1GB is not enough to fully exercise the machine.

Check out my writeread10gb script on google code.

Do the test locally on the machine, eliminating the network.

 

You can also edit the script to write only 1gb or 4gb. or the typical size of 2x you ram buffer.

 

if I were to do only a 1GB test with my tunings, I would get 50-55MB/s burst

 

Server has 8GB ram in it.

 

The disks were set as parity, so I can't test the Green again.

 

It was using AJA System Test, which is a tool that tests disks for their throughput for Video work.

 

Both disks were tested under same conditions, i.e. Green was tested, Parity was constructed on the Black disk and the tests were repeated.

 

No idea why my performance is so slow tbh.