It seems that I must install NVIDIA-linux-x86-180.22.-pkg1.run into my Linux CUDA system - even though at this time I am just emulating. I will hopefully be running native next week. Now when i install NVIDIA-linux-x86-180.22.-pkg1.run it asks me if I want to continue since my installed card is not recognized. My NVIDA card is 7 years old and I do not expect that it would be recognized as CUDA enabled. Again, at this stage I am emulating only. Since it will ignore my card should I go ahead and install. At the moment NVIDIA-linux-x86-180.22.-pkg1.run is not installed since i am emulating.
When I run some of the projects in emulation mode, I get the error CUTIL CUDA error: Kernel launch failed. in file <bisect_large.cu>, line 204 : feature is not yet implemented. The *.cu file will change depending on which project I am trying to run.
It all seems to go back to the cutilCheckMsg which it cannot find. I just want to run in emulate mode. So in order to get some of the projects to go in emulate must i install NVIDIA-linux-x86-180.22.-pkg1.run. And accept the fact that it will ignore my NVIDIA graphics card, but install some other drivers that I need.
I just cannot run some of the projects in emulate mode.
Respectfully,
newport_j