Is there a problem with the GPU call logic
Excuse me, why does the application still call the integrated graphics card of the system to work when the P100 graphics card has been installed
@luyinla2017 Please run the GUI app, go to the first screen (if it automatically loads the previous model, click the “Back” button), and click the “advanced…” button on that screen. Do you see the P100 in the last combobox?
Please run dxdiag.exe, do you have a tab for your P100 GPU?
nVidia drivers for these GPUs come in two versions. Apparently, the driver you’re using only support TCC (Tesla Compute Cluster) mode. In TCC mode, the only thing you can possibly do with the GPU is compute things with CUDA. Computing things with DirectCompute (which is what my code is doing) is not supported in that mode. Rendering 3D graphics is not supported, either.
According to nVidia, you need to pay them extra to enable Direct3D support for these GPUs. Specifically, you need to buy “virtual GPU software”, see that answer
However, if you don’t care about breaking license agreements, technically you could try the drivers intended for Google’s cloud VMs.
Note that I’m not a lawyer, and this comment is not a legal advice. Whether it’s legal to use these drivers probably depends on your jurisdiction. User-hostile EULAs like that may or may not be enforceable.
thank you