I am considering to buy a laptop to run "Tensorflow GPU" version. The MX150 2GB is not listed under the list of GPU compatible with CUDA. Does that mean it cannot run the "Tensorflow GPU" version but only the standard Tensorflow version?
Asked
Active
Viewed 9,201 times
1
-
In this [link](https://www.quora.com/Has-anyone-tried-to-run-CUDA-code-on-a-new-Nvidia-MX150-GPU) you can check the opinion about MX150 for deep learning tasks. – David Jimenez Sep 24 '18 at 09:29
2 Answers
6
Yes it can. I am currently using MX150 on my laptop with ubuntu 18.04 and I have successfully installed tensorflow gpu on it. According to MX150 on the product site it says that it is CUDA compatible. Remember to check the CUDA version you should install depending on the version of tensorflow-gpu you will be using. Good luck!
Jialer Chew
- 349
- 1
- 10
-
1Could you give us a small benchmark comparing performances on, let's say, MNIST dataset, with and without GPU ? – politinsa Apr 16 '19 at 15:26
-
0
Here is your answer : The CUDA version your should use, and example code with a comparision graph : https://vjethava.blogspot.com/2017/11/tensorflow-on-mx150.html
