- The broadcast variable is a variable that does not change inside a loop. At the start of a parfor loop, MATLAB sends the values of any broadcast variables to all the workers. The usage of large/many broadcast variables can cause significant communication between a client and its associated workers, and in turn increases overhead. You can learn more about these variables by going through the following documentation: https://www.mathworks.com/help/parallel-computing/broadcast-variable.html
- This issue arises when the loop body accesses multiple rows and columns of a matrix using the loop variable. Refer to this MATLAB Answer for more clarity upon when this issue is encountered: https://www.mathworks.com/matlabcentral/answers/419074-parfor-loop-the-entire-array-or-structure-h-is-a-broadcast-variable-this-might-result-in-unnecess
- Please note that this is a warning and not an error. There are a few ways through which you can reduce the communication overhead such as creating temporary variables within the loop body. Have a look at this MATLAB Answer which demonstrates a way to resolve the issue: https://www.mathworks.com/matlabcentral/answers/1700170-parfor-unnecessary-communication-overhead
How can I solve the unneccesary communication overhead problem caused by the broadcast variable error?
4 views (last 30 days)
Show older comments
%% Errror description Even though I tried to solve it by assigning variable Y to a temporary variable, it didn't work?
%% The entire array or structure Y-temp is broadcast variable. This might result in unneccesary comunicaiton overhead?
for k=0:1:(M-1)
z_plane_path(k+1)=A*W^(-k);
end
ChirpZ_outY(1:M)=0;
Y_temp=Y(:,1);
parfor k=0:M-1
sum1=0;
for n=0:N-1
sum1=sum1+(Y_temp(n+1)*z_plane_path(k+1)^(-n));
end
ChirpZ_outY(k+1)=sum1;
end
DFT_SS_Eq_out=ChirpZ_outY/N*2;
disp(i);
end
0 Comments
Answers (1)
Shubham
on 18 Sep 2024
Edited: Shubham
on 18 Sep 2024
Hey Alexi,
I understand that you are facing the following error: "The entire array or structure Y-temp is broadcast variable. This might result in unneccesary comunicaiton overhead" through your above code snippet.
I hope this helps!
0 Comments
See Also
Categories
Find more on Parallel for-Loops (parfor) in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!