Hi all,
I am CUDA developer and i have a Apple 15-inch MacBook Pro with Retina Display. It has 2 GPUs:
- HD Graphics 4000
- NVidia GeForce GT 650M CUDA capable.
For compile my CUDA code i have not problems but i can not run my applications. I have installed gfxCardStatus for control the GPU that is working.
The problem is:
- If i active the Nvidia GPU then it is busy with the screen graphics. I only can run applications that don't take more than 5 seconds. The time limit in the kernels is activated. This is the result of devicequery of NVIDIA sdk:
Run time limit on kernels: Yes
And my applications need more that 5 seconds. So i always get a Time out.
- If i active HD Graphics 4000 and i let free the Nvidia GPU, then it can not find any Nvidia GPU. This is the result of devicequery of NVIDIA sdk:
no CUDA-capable device is detected
So, is this MacBook CUDA capable for applications that take more that 5 seconds ?
Thanks, Moises.
I am CUDA developer and i have a Apple 15-inch MacBook Pro with Retina Display. It has 2 GPUs:
- HD Graphics 4000
- NVidia GeForce GT 650M CUDA capable.
For compile my CUDA code i have not problems but i can not run my applications. I have installed gfxCardStatus for control the GPU that is working.
The problem is:
- If i active the Nvidia GPU then it is busy with the screen graphics. I only can run applications that don't take more than 5 seconds. The time limit in the kernels is activated. This is the result of devicequery of NVIDIA sdk:
Run time limit on kernels: Yes
And my applications need more that 5 seconds. So i always get a Time out.
- If i active HD Graphics 4000 and i let free the Nvidia GPU, then it can not find any Nvidia GPU. This is the result of devicequery of NVIDIA sdk:
no CUDA-capable device is detected
So, is this MacBook CUDA capable for applications that take more that 5 seconds ?
Thanks, Moises.