Vectorization of for loop (indexing of a vector)
2 views (last 30 days)
Show older comments
Please help, I'm trying to get this code to run faster on my GPU. Currently this is the main bottleneck of my function. I was wondering whether this code is vectorizable in order to remove the for-loop? vector_1 is a gpuArray
for i=1:n
A = vector_1((i-1)+(1:10));
B = vector_1((i-1)+16+(1:10));
AA(i+10) = (A'*B)/10;
BB(i+10) = (A'*A)/10;
end
0 Comments
Accepted Answer
Joss Knight
on 18 Nov 2015
Edited: Joss Knight
on 18 Nov 2015
first = gpuArray.colon(1,n);
indexA = bsxfun(@plus, first, [0:9]');
indexB = indexA + 16;
A = vector_1(indexA);
B = vector_1(indexB);
AA(first+10) = sum(A.*B)/10;
BB(first+10) = sum(A.*A)/10;
0 Comments
More Answers (0)
See Also
Categories
Find more on Parallel Computing Fundamentals in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!