After tweaking for a couple of weeks and thanks to a lot of help from folks on this forum, I was able to finally optimize some of my testing. The main test case was trying to sort out the potential bottlenecks of the dell perc h755n. In parallel, I had a project for work that was just trying to get the most raw seq r/w throughput for some T&M sw that we make. In regards to that second effort, I was able to generate appox 300Gbps R/W sequential throughput. I realize that this is just a synthetic benchmark that is not relevant to most use cases. But some empirical data is still important at least for reference. The best case scenario was with an optimized crystaldisk test. We also did some fio tests with other systems and on different datastores. Again, thanks for the help of a lot of folks on STH that pointed me in the right direction. enjoy!