I've just browsed the iometer result file from Nizzen, random read 512B-128KB QD1-64 exp2

It looks great, a few disturbing drops in performance at high QD on some block sizes.
I'm not sure what drives he's been using, how many drives or the size of the testfile. (I expect 1GB testfile)

Access time 0.02 4KB QD1

Well, it's up to Nizzen to make this publicly available, it's well past midnight here so I expect it will be tomorrow.

I expect some graphs would be of interest