Indeed. I am aware of the diminishing returns when it comes to cache although I was a little surprised at how diminished it was.
Then again in the Norton benchmark the actual app is small and I suspect that much of the test code fit in the 256K cache and that going to 512K didn't show up in the benchmarks due to this. I will have to experiment a bit and see if there is an improvement in 'real world' cases.
Thanks again