Benchmarks for Performance Tuning and Analysis

[kennedy]

FMP and FMS performance seems to vary wildly... and tuning seems to be fairly imprecise and hit/miss... and the primary performance drivers seem to mostly be a mystery. Sooooo....

I am thinking of developing a set of benchmarks that will allow me to figure some of this out emperically. And it should allow us in this forum to run it on a wide variety of systems and setups and be able to compare performance and the impact of hardware.

To support that, I'd like to use this thread for brainstorming what ought to be in this set of benchmarks, how benchmarks should be crafted, how results should be reported, and what are the key variables that should be testable.

Also, if any performance benchmarks exist out there, please point me to them. I'd rather build on something than re-invent the wheel.

Please post any and all suggestions here.

[aklodner]

You should check with Debi Fuchs over at AptWorks. I believe she did some extensive bench mark testing and may have even written a white paper on it, but don't hold me to it. In any case, I think she'd be a great resource for this.