I would like to run several pipes in parallel, each utilizing 1/x (x = # of pipes) of the total throughput of the device. Is this possible with a threaded model, or does the library block on each call to the device? If this is true(blocks on each request), would the multi-instance model used in Linux allow me to overcome this by accessing the library/device from different instances?
I am using the 6030-LX150
Thanks,
Rich