Agree it would be nice to see something more definitive from THG on the test methodology. Sans that, and assuming THG follows many in using the patterns from Intel and defined by Storage Review...
1. File server, Workstation, Database patterns:
http://www.storagereview.com/articles/200003/20000313OSandBM_5.html
2. Web Server pattern:
http://www.storagereview.com/Testbed3.sr?page=0%2C7
3. Also see:
http://www.storagereview.com/Testbed4.sr?page=0%2C5
From the IOMeter User's Guide (assuming this is what THG is doing)...
4. Max MB/sec read/write; for read:
"For maximum throughput (Megabytes per second), try changing the Transfer Request Size to 64K, the Percent Read/Write Distribution to 100% Read, and the Percent Random/Sequential Distribution to 100% Sequential."
5. Max IOPs read/write; for read:
"For the maximum I/O rate (I/O operations per second), try changing the Transfer Request Size to 512 bytes, the Percent Read/Write Distribution to 100% Read, and the Percent Random/Sequential Distribution to 100% Sequential."
6. What isn't known about THG's tests:
a) Physical drive or file used?
- If physical, what was the maximum sector (if any)?
- If a file, how big?
b) How many workers were used?
c) The warm-up/ramp and test duration?
Hope that helps.