CudaHelpers.ixx File Reference

CUDA utility functions for device management and kernel execution. More...

#include <cuda_runtime.h>
#include <cstdint>
#include <string>
#include <format>
#include <stdexcept>
#include <cassert>
import Cuda.Error;

Namespaces
namespace	Mila
	Mila main API namespace.
namespace	Mila::Dnn
namespace	Mila::Dnn::Compute
namespace	Mila::Dnn::Compute::Cuda

Functions
constexpr int	Mila::Dnn::Compute::Cuda::ceil_div (int M, int N)
	Calculates ceiling division for kernel grid/block dimensions.
int	Mila::Dnn::Compute::Cuda::checkDevice (int deviceId)
	Validates that a device ID is valid and available.
int	Mila::Dnn::Compute::Cuda::findCudaDevice (int deviceId=-1, bool preferMemory=false)
	Finds the most appropriate CUDA device for computation.
int	Mila::Dnn::Compute::Cuda::getBestDeviceId (bool preferMemory=false)
	Identifies the best CUDA device based on performance characteristics.
int	Mila::Dnn::Compute::Cuda::getDeviceCount ()
	Gets the number of available CUDA devices.
int	Mila::Dnn::Compute::Cuda::getDriverVersion ()
	Gets the installed CUDA driver version.
int	Mila::Dnn::Compute::Cuda::getRuntimeVersion ()
	Gets the installed CUDA runtime version.
void	Mila::Dnn::Compute::Cuda::setCurrentDevice (int device_id)
	Sets the current CUDA device with thread-local caching.

Detailed Description

CUDA utility functions for device management and kernel execution.

Provides helper functions for CUDA device discovery, selection, and kernel execution optimization. These utilities simplify working with CUDA devices and provide consistent error handling.