I run the example "Stencil Operations on a GPU" in matlab command. I got following output. I don't have any advantage in using of Arrayfun on GPU compared with simple GPU array. my GPU device is GeForce GTX 980 Ti. is it reasonable?

Question

masoumeh mousavi 2016-2-21

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/269222-i-run-the-example-stencil-operations-on-a-gpu-in-matlab-command-i-got-following-output-i-don-t-h

评论： Joss Knight 2016-3-2

paralleldemo_gpu_stencil
Average time on the CPU: 1.906ms per generation.
Average time on the GPU: 0.792ms per generation (2.4x faster).
Average time using GPU arrayfun: 0.996ms per generation (1.9x faster).
CPU:          1.906ms per generation.
Simple GPU:   0.792ms per generation (2.4x faster).
Arrayfun GPU: 0.996ms per generation (1.9x faster).

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Joss Knight 2016-2-26

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/269222-i-run-the-example-stencil-operations-on-a-gpu-in-matlab-command-i-got-following-output-i-don-t-h#answer_211339

Yes, it's perfectly reasonable.

2 个评论
显示无隐藏无

masoumeh mousavi 2016-2-26

编辑：masoumeh mousavi 2016-2-26

thank you, is it the reason due to the kind of GPU? I expected more computational speed using arrayfun on GPU as claimed!(it means the usage of arrayfun strictly depends on the kind of GPU!?)

Joss Knight 2016-3-2

The speed of everything you do on the GPU depends on the kind of GPU. You should always experiment with different approaches and use whichever works best for you - we don't claim that arrayfun will always be the best approach. It looks like for your device MATLAB is doing a great job of making the simplest approach the fastest, which is ideal.

请先登录，再进行评论。