Yes. TDMS is optimized for writing not for reading, so it may be slow. But until you try it with a test case, random number array comes to mind, you do not know whether TDMS will work at all. H5(hdf) is not supported natively in LabVIEW, you will have to use one of the options in VIPM or roll your own. If you go down the road of separate files, then maybe just a flat binary structure would work.
TIP: To increase your read/write speed write/read in multiples of the disk sector size.
I think you will have to test some architectures to see what works and what hardware requirements you need.
mcduff
