speed of 6534 control using CVI (PXI-1025 w/AMI PIII equiv)

[Dave S.]

I am using 3-6534 cards to perform control/acq. I have benchmarked a for-loop and found that it takes 2mS for a port value change to take affect (no I/O, screen writing, etc. within the for-loop except the port update). What can I do to increase the speed of the system? I have bechmarked the same for-loop on an old 486-DX66 system using your competitor's PCMCIA DAQ cards and found it takes only 1.2uS to complete the same update. PLEASE HELP

[Chris Matthews]

You will get a better response for this question by posting it to the Digital I/O section of Developer's Exchange here.

Best Regards,

Chris Matthews
National Instruments

[bman]

Dave

When doing static I/O, how long it takes is very system dependant, not only processor speed, though the OS plays a big part in how long it will take.

If speed is important, I would suggest doing pattern I/O. This way, you can specify at certain rate and it will be hardware timed. There are several examples on doing Pattern I/O in the CVI >> Samples >> DAQ >> di or do directory.

Brian

[Dave S.]

Hello Brian,

Thanks for your input. I'm in a bit of a pickle. I have to do a real-time analysis of the data that's being taken and correct it if I detect an error. Using pattern I/O is not an option for me (as I see it) since the data would be taken and stored, not allowing a verification @ each location after the sample. I was shocked that the PCI-based system (purchased the 1025-PXI) I bought has about 100x slower perfomance running the same compiled-C for-loop than my old 486/PCMCIA system. I can't seem to get any support from NI on this issue...there has to be another way to increase the static I/O performance.