Dear colleagues,I'm currently trying to integrate a control system within a PXI-8106RT controller. The system is composed of the PXI-8106 controller board + 6 PXI-7813R FPGA boards. Additionally, I have a PXI-8310 card (StarFabric link), which is a kind of PCI-PCI bridge to interface with an external cRIO-9104 chassis, containing NI-9215 ADC modules, and with the cRIO-9052 StarFabric interface controller, making possible the FPGA inside the cRIO chassis to appear as another FPGA connected to the PXI bus.The PXI controller is running a loop each 500us, where the control algorithm is being executed. The control process consists of:1. Input data is read from the 6 PXI-FPGAs and from the cRIO FPGA.2. The control algorithm executes3. Outputs are provided to the 6 PXI-FPGAs.Due to stability reasons, all this 3 steps should take place in no more than 50us. Let's say that the control algorithm needs 30us to run. This means that we have 20us to read the inputs from all the 7 FPGAs and to provide the outputs to the 6 PXI FPGAs.All the data to be transmitted is:- 8 x 32 bit Unsigned Int (U32)from each PXI FPGA = 48 x U32 - 1 x 8 bit Unsigned Int (U8) from each PXI FPGA = 6 x U8- 13 x 32 bit Fixed Point (FXP) from the cRIO FPGA = 13 x 32bit FXPI have tried to use the “Read/Write control” to directly read the values from the 7 FPGAs, but it doesn’t seem to be fast enough. The results that I got after benchmarking the reading operation from the 7 FPGAs were a mean time of 126us (72us if reading only from the 6 PXI FPGAs).So I was wondering whether the timescales that I need to meet are feasible with LabVIEW real time, and which is the fastest way to transfer data between the FPGAs and the RT controller. If not, any kind of alternative would be much appreciated.Many thanks in advance,Jose
