Single page Print

IOMeter
Our IOMeter workloads feature a ramping number of concurrent I/O requests. Most desktop systems will only have a few requests in flight at any given time (87% of DriveBench 2.0 requests have a queue depth of four or less). We've extended our scaling up to 32 concurrent requests to reach the depth of the Native Command Queuing pipeline associated with the Serial ATA specification. Ramping up the number of requests also gives us a sense of how the drives might perform in more demanding enterprise environments.

We run our IOMeter tests using the fully randomized data pattern, which is a worst-case scenario for SSDs, not that it matters. They're so much faster in these IOMeter tests that including the results would completely throw off the scale of the graphs, making the mechanical drives impossible to distinguish from one another. If you want to see what I'm talking about, check out how the SSDs stack up on this page of our Vertex 4 review.

Our IOMeter workloads benefit from quick access times, and the 10k-RPM drives outclass their 7,200-RPM rivals as a result. The VelociRaptors offer much higher transaction rates from the lightest to the heaviest load. The new model extends its lead over the VR200M as the number of concurrent I/O requests ramps up in the web server, database, and workstation tests. However, the reverse is true in the file server test, where the difference between the two is particularly prominent. I suspect that's because the file server test is the only one to contain 512-byte writes, which we've already seen pose problems for the new VelociRaptor.