After my post on the Sun T1000 and the subsequent podcast interview, I noticed this comment on the T1000 post from David Halko:

Please run some tests and take the T1 to the breaking point... total number of requests, total number of failures, latency... and compare it to something else.

This is, of course, exactly what I'm doing!

I've developed a simple quite of functionality that I'll be testing. I have a comparison against Sun's own X2100 (based on the AMD) using a combination of static and dynamic elements. From that, I'll be able to extract all of the above and show the detail, as well as some graphs, on that output.