It may occur much more often on a slower machine.
We're still interested in seeing how your machine benchmarks, as Matt said above.
Here's mine (from my main "office" workstation - not sure what it all means...
=== Running Benchmarks (please do not interrupt) ===
Running 'Math' benchmark...
Single-threaded integer math... 4.690 seconds
Single-threaded floating point math... 2.924 seconds
Multi-threaded integer math... 2.421 seconds
Multi-threaded mixed math... 1.570 seconds
Score: 1637
Running 'Image' benchmark...
Image creation / destruction... 1.752 seconds
Flood filling... 1.464 seconds
Direct copying... 2.174 seconds
Small renders... 3.743 seconds
Bilinear rendering... 2.062 seconds
Bicubic rendering... 1.854 seconds
Score: 1686
Running 'Database' benchmark...
Create database... 0.916 seconds
Populate database... 2.450 seconds
Save database... 0.245 seconds
Reload database... 0.096 seconds
Search database... 2.365 seconds
Sort database... 1.721 seconds
Group database... 1.476 seconds
Score: 2320
JRMark (version 16.0.123): 1881
Cheers!
VP