I'm not a zfs guru like some folks here, but if your workloads are largely sequential, Which it sounds like they are, I'd recommend changing the record size in your pool to 1MB. You'd likely see a huge improvement.
You would have a 40gbe port on your switch configured in 4x10gbe breakout mode, and 10 gbe links to your esxi hosts.
I don't see a point in doing so unless you only have 10gbe nic in your host....
Even then, you'd be introducing a single point of failure.
I've managed a fair number of Nexus 3064 switches and I'd say they're a great switch, especially for the current price point on eBay. What are your specific requirements though? That will really drive the answer.
Traffic over a LAG is still cut-through.
Traffic from 10g to 40g can never be cut-through on any switching platform... The entire frame must be received below sending or an underrun would occur.
40g to 10g will buffer obviously, but does cut through.
As recommended, infiniband will be lower...
I've used various force 10 switches before Dell bought them, and they are excellent. The cli can take some getting used to though.
I'd say force10 is the best line that Dell has to offer in networking.