I understand,
just a little tipp. If you benchmark a storage this way you should have in mind that these values do not tell you something about what you can expect in a real environment. 100% read and 100% sequential is a very synthetic workload which you will not see in a typical virtual environment. I would recommend if you don´t know the access pattern of your typical workload that you start with a 70% read / 30% write Distribution, 80% random, 4k request size. Then you can walk your way towards a distribution which is more write dependent with different request sizes etc. and see how that works out. Vmware has a nice tool with preconfigured workloads, you find it here: