{"id":4140,"date":"2026-02-18T12:19:41","date_gmt":"2026-02-18T17:19:41","guid":{"rendered":"https:\/\/blogs.mathworks.com\/matlab\/?p=4140"},"modified":"2026-02-18T12:19:41","modified_gmt":"2026-02-18T17:19:41","slug":"faster-vector-concatenation-in-matlab-r2024b-and-later","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/matlab\/2026\/02\/18\/faster-vector-concatenation-in-matlab-r2024b-and-later\/","title":{"rendered":"Faster vector concatenation in MATLAB R2024b and later"},"content":{"rendered":"
At MathWorks, we are constantly working on making MATLAB faster. Most of the speed-ups that we achieve are discussed in the release notes but some slip through the net -- usually because they were fixed by a developer while they were in the process of working on some larger task. Here's one such optimization that found its way into MATLAB R2024b that I only recently discovered.<\/span><\/div>
Consider two vectors that you want to concatenate:<\/span><\/div>
a = [1 2 3 4];<\/span><\/span><\/div><\/div>
b = [4 5 6 7];<\/span><\/span><\/div><\/div>
c = [a b]; <\/span>% Concatenate them horizontally<\/span><\/span><\/div><\/div>
d = [a;b]; <\/span>% Concatenate them vertically<\/span><\/span><\/div><\/div><\/div>
These operations were made a little faster in R2024b and above. As you might imagine, such a small operation is very fast whatever version of MATLAB you use so we are going to repeat the concatenation 10 million times to make the point<\/span><\/div>
N = 10000000;<\/span><\/span><\/div><\/div><\/div>
% Run this in <\/span>R2024a<\/span><\/span><\/div><\/div>
tic<\/span><\/span><\/div><\/div>
for <\/span>i = 1:N<\/span><\/span><\/div><\/div>
c = [a b];<\/span><\/span><\/div><\/div>
end<\/span><\/span><\/div><\/div>
totalHorzTime = toc;<\/span><\/span><\/div><\/div>
tic<\/span><\/span><\/div><\/div>
for <\/span>i = 1:N<\/span><\/span><\/div><\/div>
d = [a;b];<\/span><\/span><\/div><\/div>
end<\/span><\/span><\/div><\/div>
totalVertTime = toc;<\/span><\/span><\/div><\/div>
 <\/div><\/div>
myver = version;<\/span><\/span><\/div><\/div>
out = <\/span>\"Ran in version \" <\/span>+ myver;<\/span><\/span><\/div><\/div>
out = out + sprintf(<\/span>\"\\nTime for %d horzcats is %.3fs\\n\"<\/span>,N,totalHorzTime);<\/span><\/span><\/div><\/div>
out = out + sprintf(<\/span>\"Time for %d vertcats is %.3fs\\n\"<\/span>,N,totalVertTime);<\/span><\/span><\/div><\/div>
disp(out);<\/span><\/span><\/div>
Ran in version 24.1.0.2689473 (R2024a) Update 6\r\nTime for 10000000 horzcats is 0.650s\r\nTime for 10000000 vertcats is 0.747s<\/div><\/div><\/div><\/div><\/div>
% Run this in <\/span>R2025b<\/span><\/span><\/div><\/div>
tic<\/span><\/span><\/div><\/div>
for <\/span>i = 1:N<\/span><\/span><\/div><\/div>
c = [a b];<\/span><\/span><\/div><\/div>
end<\/span><\/span><\/div><\/div>
totalHorzTime = toc;<\/span><\/span><\/div><\/div>
tic<\/span><\/span><\/div><\/div>
for <\/span>i = 1:N<\/span><\/span><\/div><\/div>
d = [a;b];<\/span><\/span><\/div><\/div>
end<\/span><\/span><\/div><\/div>
totalVertTime = toc;<\/span><\/span><\/div><\/div>
 <\/div><\/div>
myver = version;<\/span><\/span><\/div><\/div>
out = <\/span>\"Ran in version \" <\/span>+ myver;<\/span><\/span><\/div><\/div>
out = out + sprintf(<\/span>\"\\nTime for %d horzcats is %.3fs\\n\"<\/span>,N,totalHorzTime);<\/span><\/span><\/div><\/div>
out = out + sprintf(<\/span>\"Time for %d vertcats is %.3fs\\n\"<\/span>,N,totalVertTime);<\/span><\/span><\/div><\/div>
disp(out);<\/span><\/span><\/div>
Ran in version 25.2.0.3042426 (R2025b) Update 1\r\nTime for 10000000 horzcats is 0.557s\r\nTime for 10000000 vertcats is 0.686s<\/div><\/div><\/div><\/div><\/div>
This makes <\/span>horzcat<\/span> about 15% faster and <\/span>vertcat<\/span> around 8% faster in R2025b on this machine which, as you'll see below, is an M2 mac. On an older Intel Windows machine, I only saw a ~6% speed up in both cases so your mileage may vary.<\/span><\/div>
It's not an enormous speed-up which is probably another reason why the developer chose not to include it in the release notes. However, I thought that some of you may be interested since every little helps!<\/span><\/div>

Details of my machine<\/span><\/h2>
Using the popular <\/span>cpuinfo<\/span> add-on<\/span><\/a><\/div>
cpuinfo<\/span><\/span><\/div>
ans = struct with fields:<\/span><\/span><\/div>
CPUName: 'Apple M2 Pro'\r\n Clock: 'N\/A'\r\n Cache: 65536\r\n TotalMemory: 1.7180e+10\r\n NumCPUs: 1\r\n TotalCores: 10\r\n OSType: 'macOS'\r\n OSVersion: '15.5'\r\n Hostname: 'uk-mcrouche-m'\r\n<\/div><\/div><\/div><\/div><\/div><\/div>
<\/div>\r\n<\/div>