11-17-2009 11:24 AM
Are there any articles or posts that describe basic guidelines for choosing buffer sizes to use when acquiring data?
Are there specific problems that tend to occur when buffers are either too large or too small?
Thanks,
Dave
Solved! Go to Solution.
11-17-2009 11:31 AM
The buffer has to be large enough to handle the data acquired durring the longest interval between reading th backlog.
Too small and you run into issues if you don't read your data fast enough.
Too large and you waste memory.
My personal guideline is "close my eyes and imagine the user trying to open a huge Excell spreadsheet while the app is running"- however long I suspect the PC will take to open the file will dictate (based on sample rate) what is the minimum size. In the old-days I would just double that number. These days i can use a fudge factor of 5.
Ben
11-17-2009 12:17 PM
Thanks. Is the backlog the data which has been buffered but has not yet been read?
11-17-2009 12:22 PM
davey31415 wrote:Thanks. Is the backlog the data which has been buffered but has not yet been read?
Yes. i believe it used to be called the "backlog" in the old traditional DAQ VIs. In DAQmx there is a property that report the number of samples waiting in the buffer.
Ben