Enum CudaDeviceRuntimeFlags
CUDA runtime device scheduling and mapped-host-memory flags. CUDA runtime 设备调度与 mapped host memory 标志。
[Flags]
public enum CudaDeviceRuntimeFlags : uint
Fields
LocalMemoryResizeToMax = 16Keep local memory allocations after launch. launch 后保留 local memory 分配。
MapHost = 8Enable mapped host memory support. 启用 mapped host memory 支持。
ScheduleAuto = 0Let CUDA choose the device scheduling behavior automatically. 让 CUDA 自动选择设备调度行为。
ScheduleBlockingSync = 4Use blocking synchronization while waiting for device work. 等待设备工作时使用阻塞同步。
ScheduleSpin = 1Spin while waiting for device work to complete. 等待设备工作完成时使用自旋。
ScheduleYield = 2Yield the CPU thread while waiting for device work. 等待设备工作时让出 CPU 线程。