vectorization examples for efficiency

1 次查看（过去 30 天）
feynman feynman
feynman feynman 2024-3-10
评论： feynman feynman 2024-3-11
I wonder if the following expressions are already optimized forms for computational efficiency and if not how to rewrite them?
x=linspace(0,1);
0*x; x+1
a=[x;x]
[2*x;a(1,:)]
repmat([1;1;0],1,10);

请先登录，再回答此问题。

采纳的回答

Walter Roberson
Walter Roberson 2024-3-10
x = 0:1/99:1;
zeros(size(x)); x+1
a=[x;x];
[2*x; x]
repmat([1;1;0],1,10);
  3 个评论
显示 1更早的评论
Walter Roberson
Walter Roberson 2024-3-11
编辑：Walter Roberson 2024-3-11
Ran in:
Interesting
When I try it several times, the times vary pretty wildly, including cases where the 0*x comes out much slower.
format long g
testit();
testit();
T = testit()
T = 3×1
0.000366 0.001324 0.000718
function T = testit()
T = zeros(3,1);
N = 10000;
x = linspace(0,1);
start = tic; for K = 1:N; Z = 0*x; end; T(1) = toc(start);
start = tic; for K = 1:N; Z = zeros(size(x)); end; T(2) = toc(start);
start = tic; for K = 1:N; Z = zeros(1,100); end; T(3) = toc(start);
end
feynman feynman
feynman feynman 2024-3-11
, which means 0*x and zeros(size(x)) aren't necessarily better or worse than the other?

请先登录，再进行评论。

更多回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Characters and Strings 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by