Class providing information about the CUDA devices at runtime.
Without exposing any CUDA dependencies publicly to the clients.
Declarations preventing any copies of the singleton object
Helper struct describing one available CUDA device.
bool concurrentKernels = false¶
Whether the device supports multiple kernel executions in parallel.
int id = -1¶
Identifier that CUDA knows this device by.
int maxThreadsPerBlock = -1¶
The maximal number of threads per block for this device.
The name of this device.
std::size_t totalMemory = 0¶
The total amount of (global) memory on the device.
- bool concurrentKernels = false¶
- struct Device¶