First of all, my apologizes, due to it is possible that this question may be a easy question. I´ve seen some messages about this, but i didn´t found exactly what i need.
I am developing a very "busy" I/O system. I need, each 40 ms, generate and acquired 680k words (32-bit wide). Rate of PCI bus (2.0, its our choise) is more fast than my system requirements (2 or 3 more fast). My system signals which shall be captured takes about 28 - 29 ms (the 680k words), and the amount of data (same words) over the PCI bus takes 11ms (without any more considerations like latency, addressing, etc, only data). I am thinking on send data to generate 12 ms before the first 40ms period, and after that, for each period, send data on the last 12ms of the full period.
For adquisition, due to it is the same amount of data, I am thinking on use the first 12 ms period (every period) to send captured data to the Host. I have seen the niHSDIO_WaitUntilDone function, so, my questions are: Can I initiate a dynamic adquisition season (taking into account samples, etc), which needs at least 29ms to capture all, while it continues using the PCI bus for other devices, using niHSDIO_WaitUntilDone function? Which is the signal that indicates to the Host that the adquisition have finished? Does It take the bus all the time?
Thanks
Mensaje editado por MMjj
