NVIDIA Geforce MX150 support cuda or not

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

HI Team,

We want to purchase a 13-14 a laptop for AI Learning that support CUDA.I checked the laptop and many laptop has NVIDIA Geforce MX150 card on it , while going through forum i saw that user has faced issue with cuda with NVIDIA Geforce MX150 graphic card but on your link it said NVIDIA Geforce MX150 support cuda. I am little bit confused so please tell me whether we should NVIDIA Geforce MX150 graphic card or not as laptop is very costly.

2

laptop we are looking for 13 to 14 inch size

3

we want run cuda with tensorflow and gpu so please suggest

4

HI plz reply on above thread

5

GeForce MX150 supports CUDA.

6

Does it support high end gpu applicaton

7

It has various limitations, such as memory size.

8

I’ve installed tensor rt 8.2 with cuda-toolkit 10.2 on geforce mx150.
But running some functions of tensor rt, it cause oom.
Even though I changed the argument of IBuilder::setMaxWorkspaceSize , it didn’t work.
Any suggestion or any advice for the issue?
Please…
Should I install tensor rt 7.x ? Is that able to solve this issue?

13

It’s entirely possible that attempting to serialize and use a large model in TensorRT would fail due to out of memory. To perform inferencing on a model, TensorRT needs to be able to load the model into GPU memory. If the model is large enough, a GPU with very small memory size like MX150 might not be able to hold it.

I doubt an earlier (or other) version of TensorRT would help here, if model size is the issue. The recommendation at that point would be to use a GPU with more memory.