I am taking 40 channels of AI while simultaneously doing 3 buffered
counter operations.  The 40 AI channels are coming from PXI 6608's
(which have plenty of onboard memory to handle the acquisitions).

The problem is that the 3 counter DMA transfers are hampered by the AI
channels reading the onboard memory across the PCI bus, so I get very
low bandwidth on the counters.

How can I define the counter DMA transfers to have priority over the
AI transfers?  I want the AI data to just stay in the onboard memory
until the counter DMA is finished, then transfer to the PC memory
buffer.  That way, the PCI bandwidth is being used where it is needed.

Thanks,

rgames

Reply via email to