readout fast large data

[commander]

hi i have a question:

i have a lage file and what i need to readout this file word by word. so the code is working, but very slow. how can i readout it faster, what is the best way, because i need this data  pixel by pixe (each pixel is a 32 bit vektor) the pixel are safed in a 1D array.

So one idea: two processes. first process:readout the file in array by array mode (not pixel by pixel) buffer this array in a queue and in a second process i read this array pixel by pixel out.of this queue and the first process reads the next 1 d array from the file... is this a good idea?

THX

[ramses64]

Hi,

just a trivial answer : remove the "wait 1000ms, but I'm quite sure that you already thought about it...
Best regards,
Vincent 

[Coq_Rouge]

Why are you reading the file byte by byte? It is much faster to read all the bytes in the file in one operation. You are checking the file size every time your loop runs. Is this really necessary? I have tuned your VI test the difference

[commander]

yes it is necessary to read the file byte by byte, because i have to select something out, and to convert the 13 bit vectors.

THX for the idea with the 100 ms. but this is clear:)

[commander]

Why are you reading the file byte by byte? It is much faster to read all the bytes in the file in one operation. You are checking the file size every time your loop runs. Is this really necessary? I have tuned your VI test the difference

I need a pointer, to know what is the nect bzte i have to readout...or not?

[Coq_Rouge]

The file pointer remember current pos. So you do not have to bother with it.

[commander]

ok thx that's the first advancement, but it is also to slow