how to reduce a processing time in a FOR LOOP

3 次查看（过去 30 天）

Luiz Bernardo 2012-12-26

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/57447-how-to-reduce-a-processing-time-in-a-for-loop

Hi, I have the following problem. I'm simulating a ODFM system that I must create a BERxSNR graphic using a for command . The problem is that I need to apply a viterbi algorithm two times inside this loop to estimate the signal. But this takes a long time to process (some times several days). I would like to have an option to reduce this time , below follows a very basic sample of the code

for isnr =2:1:30 (% SNR)

      received signal to be analysed; %that is a [matrix 1 x (M x  N)]
       funtion viterbi 1 %( Its a loop point by point)
         real(recived signal to be analysed)
          output 1 ( [matrix 1 x (M/2 x  N/2)]
       funtion viterbi 2 ( Its a loop point by point)
       imag(recived signal to be analysed)
       output 2 ( [matrix 1 x (M/2 x  N/2)] 
      data estimated=([output 1, output 2],MxN,1);% recompose the estimate data

end

[no_of_error(isnr),ratioisnr)]=biterr(data in , data_estimated)

2 个评论
显示无隐藏无

Walter Roberson 2012-12-26

Are your two functions really defined inside the "for" loop ??

Your outline does not show "isnr" being used?

Jan 2012-12-27

Does the profiler reveal, which lines require the most time? What are usual sizes of "M" and "N"?

It is most likely impossible to give a specific suggestion for an improvement based on an outlines code only. Posting the relevant (most time consuming) part of the real code would be more helpful.

请先登录，再进行评论。

请先登录，再回答此问题。