How does one use the GPU without switching off Optimus?
There has been discussion of using machines like the MacBook Pro for CUDA computing. The SDK samples fail unless you install a third party switching application to force Optimus to use the NVidia GPU.
The CUDA Optimus developer’s guide includes code samples that I found to again fail if the machine is in its normal dynamic switching mode.
This would seem to pose a problem for distributing software such that it does not require that the end user manually switch modes.