Hi

[ElbaTech]

does the memory buffer of DIG_Block_In() have any restriction ? NI suggests to use double buffering when the buffer is large. However is this a must or an option ? In addition, if the target buffer is "small", say around 1000 data, does it exist a preference for its dimension (i.e. power of 2, a number depending on the page size, ...) ? Does all of the above change with the OS ? (I am working with an AT-DIO-32HS under LW CVI 5.5.1)

[Nick-W]

Elba,
The restrictions on the buffer size are given in the NI-DAQ Online Help descriptions for DIG_Block_In. Basically, the size is based on your number of samples and channels. You should use double buffering is you want to do continuous acquisiton. If you just want to take a single buffer of data, you don't have to use Double buffering. The advantage of using double buffering instead of a single large buffer is that you can do assynchronous operations, instead of synchronous.
OS only come into play on NT/2000 when you get to buffers of 64MB.

Nick Wilson
National Instruments
www.ni.com/ask