Using fread to read streaming binary data from ethernet using Instrument Control Toolbox

Question

0 个投票

My data are uint8 format. The documentation for fread seems to imply that the "precision" argument can set the output format of fread. But if I try to use precision='*uint8' or 'uint8=>uint8' I get an error. The default output is double, which takes too much memory. I've been using

Output=uint8(fread(TcpipObj,SIZE,'uint8'))

which works, but I'm guessing this generates a large array of doubles then converts to uint8.

My instrument streams data until the buffer is full, so I set up TcpipObj.InputBufferSize to pull in the data I need. I then set SIZE to the same value. I'd like to pull in 40 MB or so, but get a java heap memory error.

Suggestions?

Jim

4 个评论
显示 2更早的评论隐藏 2更早的评论

James Watson 2012-4-24

I seem now to have a hardware problem that's preventing me from reproducing the error and giving you a look at it. I'll try to reply when I get that squared away.

James Watson 2012-7-18

My instrument is back from repair! The error I get is Java exception occurred: java.lang.OutOfMemoryError: Java heap space at com.mathworks.toolbox.instrument.InputBuffer.getDataFromBuffer(InputBuffer.java:705) at com.mathworks.toolbox.instrument.InputBuffer.getDataWithInfo(InputBuffer.java:528) at . . . plus several more lines that I can type out if anyone is interested.

It appears that fread in basic Matlab and fread in the instrument toolbox are different functions with different input variables allowed.

请先登录，再进行评论。

请先登录，再回答此问题。

Follow Question