3 image:performance-chart.png[alt="CUDA/Ada performance"]
5 The chart shows the cumulated execution times of performing a matrix
6 multiplication operation on a 512 by 512 matrix 20 times. All CUDA
7 implementations used the same kernel, a grid size of 32 and a block size of 16.
12 ---------------------------------------------------------------------
13 include::../src/add.adb[]
14 ---------------------------------------------------------------------