How to speed up my code

f= 'test_new.xlsx';
for i=2687:6324
    k=num2str(i);
    x=num2str(i+1);
    xlsrange=['A',x];
    FileNames=['acc (',k,').46']
    Data=dlmread(FileNames);
  j=3;
    M=0;
    M1=rms(Data(:,j));                      %% rms
    M2=kurtosis(Data(:,j));                 %% kurtosis
    M3=max(abs(Data(:,j)));                 %% peak
    M4=max(abs(Data(:,j)))./rms(Data(:,1)); %% crest factor
    M5=std(Data(:,j));                      %% Standard deviation 
    M6=rssq(Data(:,j));                     %% Root-sum-of-squares level
    M7=var(Data(:,j));                      %% Variance
    M8=peak2peak(Data(:,j));                %% Maximum-to-minimum difference
    M9=min(Data(:,j));       %% Smallest elements in array
    M10=max(Data(:,j));      %% Largest elements in array
    M11=peak2rms(Data(:,j)); %% Peak-magnitude-to-RMS ratio
    M12=meanfreq(Data(:,j)); %% Mean frequency
    M13=medfreq(Data(:,j));  %% Median frequency
    M14=range(Data(:,j));    %% Range of values
    M15=mad(Data(:,j));      %%  Mean or median absolute deviation
    M16=iqr(Data(:,j));      %% Interquartile range
    M17=harmmean(Data(:,j)); %% Harmonic mean
    M18=snr(Data(:,j));      %% Signal-to-noise ratio
    M19=powerbw(Data(:,j));  %% Power bandwidth
    M20=enbw(Data(:,j));     %% Equivalent noise bandwidth
    M21=obw(Data(:,j));      %% Occupied bandwidth
    M22=sfdr(Data(:,j));     %% Spurious free dynamic range
    M23=sinad(Data(:,j));    %% Signal to noise and distortion ratio
    M24=thd(Data(:,j));      %% Total harmonic distortion
    M25=toi(Data(:,j));      %% Third-order intercept point
    M26=median(Data(:,j));   %% median
    M27=mean(Data(:,j));     %% mean
    M28=mode(Data(:,j));     %% Mod
    time=10*(i);
    M=horzcat(time,M1,M2,M3,M4,M5,M6,M7,M8,M9,M10,M11,M12,M13,M14,M15,M16,M17,M18,M19,M20,M21,M22,M23,M24,M25,M26,M27,M28);
    xlswrite(f,M,j,xlsrange);
end

My code reads 6000 files ,each file including 20k data rows in one column.How can i make this run faster.Its taking around 3 hours now.

Answer 1

[KSSV]

I think almost all the codes up there are vectorised......you can striaght away use a matrix and specify the dimension i.e you want calculation along row or column. Please read the documentation. YOu need not to run a loop, you can striaght away use the matrix as input and specify the dimension.

Ex: RMS

A = rand(100) ;
rms(A,1)   % rms along columns 
rms(A,2)  % rms along rows

[Tanmay Agarwal]

Its a fact that vectorized code is faster in matlab,suggest something else

[Tanmay Agarwal]

And Data in this code is a 20480x4 matrix